Citation
Pregnant Female Anthropomorphic Dosimetry in Computed Tomography

Material Information

Title:
Pregnant Female Anthropomorphic Dosimetry in Computed Tomography
Creator:
Hoerner, Matthew R
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (238 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Biomedical Engineering
Committee Chair:
HINTENLANG,DAVID ERIC
Committee Co-Chair:
BOLCH,WESLEY EMMETT
Committee Members:
PROPERZIO,WILLIAM S
BACIAK,JAMES EDWARD
Graduation Date:
5/3/2014

Subjects

Subjects / Keywords:
Dead time ( jstor )
Dosage ( jstor )
Dosimetry ( jstor )
Fetus ( jstor )
Ionization chambers ( jstor )
Optical fibers ( jstor )
Photons ( jstor )
Predetermined motion time systems ( jstor )
Scintillation counters ( jstor )
Solar X rays ( jstor )
Biomedical Engineering -- Dissertations, Academic -- UF
ct -- dosimetry -- fetal -- fiber-optic -- scintillator
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Biomedical Engineering thesis, Ph.D.

Notes

Abstract:
Over the last twenty five years, medical imaging utilizing ionizing radiation has significantly increased and as a result an individual is more susceptible to receive a higher exposure over the course of his or her lifetime. Pregnant female diagnostic studies have followed this trend as there has been a 25% increase in Computed Tomography (CT) imaging studies from 1997 to 2006. An increase in the number studies coupled with newer scanner technology, which has shown to maintain if not raise radiation exposure for a single exam, have resulted in significant increase in individual and collective effective doses. Most pregnant women that undergo CT exams will receive less than 50 mGy over the course of their pregnancy. This level of absorbed dose is the established threshold for concern, as it is less than what is needed to induce abnormalities such as growth and mental retardation, which are common effects documented in the epidemiological studies of the Japanese bomb survivors. However, this level of exposure is comparable to radiation workers who by law require individual monitoring. The regulation limit set by the Nuclear Regulatory Commission (NRC) is established that no person shall receive more than 50 mSv of radiation dose in a given year. For an occupational worker who is declared pregnant, the limit is 5mSv per pregnancy cycle or 0.5mSv in one any month. Given the sensitivity of young children to radiation compared to adults, this discrepancy in monitoring needs to be addressed in some form. Using anatomically accurate reference physical models and recently developed advanced dosimetry techniques, the absorbed dose to the fetus and embryo will be evaluated at multiple gestation stages from clinical CT studies performed at the conventional imaging protocols using a modern scanner. The results of these studies are compared to both current biological data to validate or disprove claims about radiation safety in diagnostic radiology, and current clinical fetal and embryonic dosimetry methods. Most of these resolved uncertainties stem from the effectiveness of tube current modulation imaging and the accuracy of dose calculations from scanner reported DLP values, especially for cases when the beam width is greater than 10cm. The manuscript is also intended to advocate any additional imaging precautions, or necessitate new and advanced techniques of monitoring conceptus doses from clinical CT studies. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2014.
Local:
Adviser: HINTENLANG,DAVID ERIC.
Local:
Co-adviser: BOLCH,WESLEY EMMETT.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2015-05-31
Statement of Responsibility:
by Matthew R Hoerner.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
5/31/2015
Classification:
LD1780 2014 ( lcc )

Downloads

This item has the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EIKTI9J0O_P5QDSW INGEST_TIME 2014-10-03T22:27:10Z PACKAGE UFE0046653_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

PREGNANT FEMALE ANTHROPOMORPHIC DOSIMETRY IN COMPUTED TOMOGRAPHY By MATTHEW R HOERNER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014

PAGE 2

2014 M atthew R. Hoerner

PAGE 3

To my family Mom, Dad, and Mike

PAGE 4

4 ACKNOWLEDGMENTS First and foremost I would like to thank my family. Without their unconditional love, guidance, and financial support grad school would likely not of been possible. They instilled in me the work ethic, drive, passion, and character that allowed me to aim high and reach all the goals I have accomplished. It is these reasons why I have dedicated this dissertation to them. I would also like to thank those previous students who have mentored me over the course of my time as a graduate research assistant at UF This includes, but not limited to: Chris Tien, Dan Hyer, Ryan Fisher, Matt Maynard, and Perry Johnson. I am now proud and honored to be your colleague. Y our knowledge, passion for study, and guidance, has paved the w ay for me to excel in this program just as you once did. You are all what I strive to be and I am certainly proud to call you colleague, but you will always be above all, friends. Just as important I must acknowledge some great friends and fellow resear chers : David Borrego, Elliott Stepusin, Danny Long, Amy Geyer, Anna Mench, and Shannon OReily Knowing them has been a huge blessing and made my grad school experience wonderful. I look forward to future collaborations as well as watching and supporting you through the doctoral process. Most importantly, I want to thank Dr. Hintenlang. You have been an amazing advisor and I am forever indebted to you for the knowledge you have imparted on me and giving me th e opportunity to be your student. You have answered every email, always kept your door open, and dealt with my crazy antics. I look forward to continuing to annoy you for advice and guidance as I begin my career. In addition to Dr. Hintenlang, I also want to thank my entire doctoral committee. Giving up your valuable time to serve on my committee is commendable and very much appreciated.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 11 ABSTRACT ................................................................................................................... 14 CHAPTER 1 SUMMARY OF ESTIMATING CONCEPTUS DOSE, EVALUATING STOCHASTIC RISK, AND THE EFFECTS OF RADIATION ON UNBORN CHILDREN ............................................................................................................. 16 Clinical Dosimetry and Risk .................................................................................... 16 Clinical CT Dosimetry ....................................................................................... 17 Clinical Fluoroscopy Dosimetry ........................................................................ 21 General Risk Assessment ................................................................................ 22 BIER VII ............................................................................................................ 24 Studies Evaluating Conceptus Dose ................................................................ 25 Radiobiology ........................................................................................................... 27 Gestational Periods .......................................................................................... 28 Preimplantation .......................................................................................... 28 Organogenesis ........................................................................................... 29 Fetogenesis ............................................................................................... 29 Specific Developmental and Carcinogenetic Effects ........................................ 30 ICRP 90 Conclusions ....................................................................................... 32 Summary of E ffects ................................................................................................ 34 Research Objective ................................................................................................ 36 2 HIGH ENERGY COUNT RATE EFFECTS AND BACKGROUND REMOVALISSUES IN PLASTIC SCINTILLATING FIBER OPTIC DOSIMETRY ... 40 Counting Issues in Fiber Optic Dosimeters Used in Radiotherapy Dosimetry ........ 40 The Stem Effect ...................................................................................................... 41 Cerenkov Radiation .......................................................................................... 41 Color Centers ................................................................................................... 43 Ideal Stem Respons e to Radiation for a Pair of Optical Fibers ......................... 45 Count Rate Effects .................................................................................................. 47 Radiation Delivery Mechanisms of Clinical Radiation Generators .................... 47 Continuous sources ................................................................................... 47 Linear accelerator (LINAC) ........................................................................ 47 Dead Time Effects from a PMT Detection System ........................................... 48

PAGE 6

6 Dead Time Modeling ........................................................................................ 49 Materials and Methods ............................................................................................ 51 Detection System ............................................................................................. 51 Sources of Radiation ........................................................................................ 53 Experimental Methods ...................................................................................... 54 Stem count behavior .................................................................................. 54 System time resolution ............................................................................... 54 Gadolinium oxysulfide and the plastic performance ................................... 55 Percent depth dose .................................................................................... 56 Results .................................................................................................................... 57 Gadolinium Oxysulfide Scintillator vs. the Plastic Scintillator ............................ 57 Stem Count Behavior ....................................................................................... 58 System Time Resolution ................................................................................... 58 Continuous sources ................................................................................... 58 Pulsed x ray source ................................................................................... 59 Percent Dept h Dose Curves ............................................................................. 6 0 Random Phenomena ....................................................................................... 60 Discussion .............................................................................................................. 61 Conclusion .............................................................................................................. 63 3 DEVELOPMENT OF AN ENGERY INDEPENDENT FIBER OPTIC DOSIMETER FOR RADIOLOGICAL DOSIMETRY ................................................ 76 Energy Dependence in Radiological Dosimetry ...................................................... 76 Thermoluminescent Dosimeters ....................................................................... 78 Optically Stimulated Luminescent Dosimeters ................................................. 78 Metal Oxide Semiconductor FieldEffect Transistor ......................................... 79 Tissue Equivalent Fiber Optic Dosimeter ................................................................ 80 Background ...................................................................................................... 80 Two Scintillators Compacted into a Single Dosimeter P ackage ....................... 81 Materials and Methods ............................................................................................ 82 Construction of the Dosimeter .......................................................................... 82 Determining the System Time Resolution and Scintillation Efficiency .............. 84 Exposure Measurements .................................................................................. 86 Free in ai r measurements .......................................................................... 86 CTDI phantom measurements ................................................................... 86 Angular dependence .................................................................................. 87 Calculation of error ..................................................................................... 88 Monte Carlo Methods ....................................................................................... 89 Generating photon and electron energy fluence data ................................ 89 Evaluation of ionization quenching from Birks model ................................ 91 Results .................................................................................................................... 93 Count Rate Performance .................................................................................. 93 Monte Carlo ...................................................................................................... 94 Energy Response ............................................................................................. 95 Angular Dependence ........................................................................................ 97 Discussion .............................................................................................................. 97

PAGE 7

7 Conclusion ............................................................................................................ 100 4 CHARACTERIZATION OF A LINEAR FIBEROPTIC DOSIMETER FOR COMPUTED TOMOGRAPHY .............................................................................. 111 Introduction ........................................................................................................... 111 Fiber Optic Dosime ters ................................................................................... 112 Plastic Scintillating Dosimeters ....................................................................... 114 Materials and Methods .......................................................................................... 114 Simulated Spatial Sensitivity .......................................................................... 114 Construction ................................................................................................... 116 Characterization ............................................................................................. 117 General characterization .......................................................................... 117 Characterization for computed tomography dosimetry ............................. 119 Results .................................................................................................................. 121 Computational Simulations ............................................................................. 121 Linearity and Reproducibility .......................................................................... 122 Energy Dependence ....................................................................................... 122 Spatial Sensit ivity ........................................................................................... 123 Characterization for Computed Tomography Dosimetry ................................ 124 Discussion ............................................................................................................ 126 Conclusion ............................................................................................................ 127 5 CONTRUCTION OF AN ANTHROPOMORPHIC PREGNANT FEMALE PHANTOM SERIES FOR COMPUTED TOMOGRAPHY DOSIMETRY ............... 141 Physical Phantoms for Experimental Radiation Dosimetry ................................... 141 University of Florida Physical and Computational Phantom Series ................ 142 Research Based Pregnant Female Anthropomorphic Phantoms ................... 143 Building a Novel Anthropomorphic Phantom .................................................. 143 Materials and Methods .......................................................................................... 144 Measurement Techniques .............................................................................. 144 Hydrostatic weighting technique .............................................................. 144 Hounsfield number ................................................................................... 146 Linear attenuation coefficient ................................................................... 146 Materials ......................................................................................................... 149 Soft tissue equivalent substitute (STES) .................................................. 149 Adipose tissueequivalent substitute (ATES) ........................................... 149 Bone tissue equivalent substitute (BTES) ................................................ 150 Phantom Construction Methodology .............................................................. 151 Design and development of anatomy ....................................................... 151 Selection of fetal anatomy for dosimetry .................................................. 152 Fabrication using a 3D printer .................................................................. 153 Results .................................................................................................................. 154 Materials and Measurements ......................................................................... 154 Comp leted Set of Phantoms ........................................................................... 155 Discussion ............................................................................................................ 156

PAGE 8

8 Conclusion ............................................................................................................ 157 6 MEASURED CONCEPTUS DOSES ON AN ANTHROPOMORPHIC PHANTOM FROM MULTI DETECTOR COMPUTED TOMOGRAPHY STUDIES USING TUBE CURRENT MODULATION FOR ALL GESTATIONAL STAGE S ................ 165 Introduction ........................................................................................................... 165 Materials and Methods .......................................................................................... 168 Anthropomorphic Phantoms ........................................................................... 168 Measurement Techniques .............................................................................. 169 PMT Control Program .................................................................................... 170 CT Procedures ............................................................................................... 171 Assessing Dose to the Fetus .......................................................................... 171 Results .................................................................................................................. 172 Absorbed Dose Measurements ...................................................................... 172 Dose Estimation Models ................................................................................. 173 Discussion ............................................................................................................ 174 Conclusion ............................................................................................................ 177 7 CONCLUSIONS AND FUTURE WORK ............................................................... 186 Plastic Fiber Optic Detectors ................................................................................ 186 General Notes ................................................................................................ 186 Energy Dependence of a Plastic Scintillator ................................................... 187 Anthropomorphic Phantoms ................................................................................. 189 Physical Dosimetry ............................................................................................... 189 Future Work .......................................................................................................... 191 Future Work Project A .................................................................................... 191 Future Wor k Project B .................................................................................... 191 Future Work Project C .................................................................................... 192 APPENDIX A MATLAB CODE FOR GRAPHICAL INTERFACE ................................................. 196 B DERIVATION FOR WEIGHTING LINEAR FIBER OPTIC DOSIMETER MEASUREMENTS ............................................................................................... 2 21 LIST OF REFERENCES ............................................................................................. 226 BIOGRAPHICAL SKETCH .......................................................................................... 238

PAGE 9

9 LIST OF TABLES Table page 1 1 Comparison of effective dose calculated from CT dose software for female from multidetector CT (MDCT) ........................................................................... 38 1 2 Conceptus doses from a sample of common clinical CT studies ........................ 38 1 3 Lifetime risk from a 50 mGy acute exposure to a 01 year old child. .................. 39 2 1 Measured pulse width at half maximum and pulse period for the Varian Trilogy accelerator .............................................................................................. 65 2 2 Plastic counts from signal and reference fibers as well as output factors. .......... 65 2 3 Gadolinium counts from signal and reference fibers as well as output factors ... 65 2 4 Calculated dead times from various radiation sources ....................................... 66 2 5 Calculated dead time parameters of a 6 MV photon beam freein air and percent depth dose measurements. ................................................................... 66 3 1 Common properties of GOS and polystryene ................................................... 102 3 2 Polystyrene dead time and correction factor evaluation from freein air measurements and Monte Carlo methods ........................................................ 102 3 3 GOS dead time and correction factor evaluation from freein air measurements and Monte Carlo methods ........................................................ 103 3 4 Response factor comparison between estimated air kerma conversion coefficient vs. measured. ............................................................................... 103 4 1 Optical properties of the plastic scintillator and optical waveguide ................... 129 4 2 Reproducibility and linearity measurements with 7cm and 15cm linear dosimeters ........................................................................................................ 130 4 3 Beam profile data from the full body phantom scan at 100, 120, and 140 kV. .. 130 5 1 Commercially available anthropomorphic phantoms used for research ........... 158 6 1 Dose coefficients for converting fetal soft tissue doses to total body doses. .... 179 6 2 Imaging parameters for 16MDCT scanner. ..................................................... 179 6 3 Radiation absorbed embryo doses from a variety of pelvic exams. .................. 180

PAGE 10

10 6 4 Radiation absorbed 15 week fetal doses from a variety of pelvic exams. ........ 180 6 5 Radiation absorbed 25 week fetal doses from a variety of pelvic exams. ........ 181 6 6 Radiation absorbed 38 week fetal doses from a variety of pelvic exams. ........ 181

PAGE 11

11 LIST OF FIGURES Figure page 2 1 Schematic illustrating the acceptance cone of the optical fiber in relation to the Cerenkov emission cone of the electron transversing through the fiber. ...... 67 2 2 Pulsing kinetics of a typical LINAC. .................................................................... 67 2 3 PMT time response from a finite light pulse at the photocathode. ...................... 68 2 4 Pulse repetition rate effects on the dosimeters response. ................................ 69 2 5 Optical fiber response to varying lengths of radiation exposures delivered in a series of field sizes to the optical fiber. ............................................................... 70 2 6 Plot of inverse count rate and dose rate to measure dead times from cobalt 60, orthovoltage, and clinical protons ................................................................. 71 2 7 Megavoltage beam irradiation freein air ............................................................ 72 2 8 Percent depth dose curves for different field sizes. ............................................ 73 2 9 Spontaneous pulse from the detection system. .................................................. 74 2 10 Counting profiles from incremental insertion of the PSD into the cobalt irradiator ............................................................................................................. 75 3 1 Size of the dual scintillator dosimeter relative to a United States quarter. ........ 104 3 2 Count rate effects from 120keV x ray exposure to the GOS scintillator and the plastic scintillator ........................................................................................ 105 3 3 Electron fluence spectrum through the polystyrene scintillator for various hole locations in the CTDI phantom from exposure to a CT x ray beam .................. 106 3 4 Gadolinium oxysulfide and polystyrene relative photon fluence spectrum from a 120kVp CT x ray beam. ............................................................................... 107 3 5 Gadolinium oxysulfide and plastic scintillator energy response as a function of depth into CTDI phantom. ............................................................................. 108 3 6 D etermination of correction factor for the plastic scintillator for energies. ......... 109 3 7 A ngular dependence of the response ratio ....................................................... 110 4 1 I llustration of the coupling efficiency analysis. .................................................. 131 4 2 T issue equivalent phantoms for depth dose measurements ............................ 132

PAGE 12

12 4 3 P hantom assembly and centered LFOD moving together from the right to left over the scanning length. ................................................................................. 133 4 4 R elative transmission of light through the optical scintillating fiber boundary at locations along the scintillating fibers axis ................................................... 134 4 5 R elative transmission of light through the optical scintillating fiber boundary at locations along the scintillating fibers axis ................................................... 135 4 6 L FOD response factor measured for varying x ray tube voltage ...................... 136 4 7 L FOD response factor measured for varying penetration depths into a soft tissue equivalent medium ................................................................................. 136 4 8 E valuation of the 7cm, 12cm, 10 cm, and 15cm LFOD dosimeters for spatial sensitivity using a uniform intensity x ray field ....................................... 137 4 9 L FOD response to varying collimation widths ................................................... 138 4 10 Distribution of counts from discrete time intervals sampled over a helical CT irradiation of a 10cm LFOD in a custom CTDI phantom .................................. 139 4 11 Ga ussian fits of each counting profile from (a) LFOD and (b) GSD. ................. 140 5 1 F rontal view of the pregnant computational hybrid phantoms pelvis. ............... 159 5 2 Se lection of cutting planes to divide the abdomen/pelvis of each phantom. ... 160 5 3 Pr ogression of mold fabrication ........................................................................ 161 5 4 De nsity curve for ATES based on added weigh percentage of micobeads. ..... 162 5 5 P regnant female anthropomorphic abdomen/pelvis phantom series ................ 163 5 6 Topogram of the 25week phantom .................................................................. 164 6 1 Phantom assembly used to calibrate the dosimeters ....................................... 182 6 2 MATLAB GUI dosimetry program with user functions and output data. ............ 183 6 3 Relationship between fetal dose and CTDIvol at various gestational ages for a CAP examination. .......................................................................................... 184 6 4 Accumulated fetal dose as function of scan position ........................................ 185 7 1 Plot of / ratio between polystyrene and air in the low energy range. ........ 194 7 2 Plot of / ratio between various dosimeters and air in the low energy range. ............................................................................................................... 195

PAGE 13

13 B 1 Dosimeter alignment for performing organ dosimetry. ...................................... 221 B 2 Geometry of an ellipse. ..................................................................................... 222 B 3 Dosimeters aligning to geometry of an ellipse. ................................................. 224

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PREGNANT FEMALE ANTHROPOMORPHIC DOSIMETRY IN COMPUTED TOMOGRAPHY By Matthew R Hoerner May 2014 Chair: David E Hintenlang Major: Biomedical Engineering Over the last twenty five years, medical imaging utilizing ionizing radiation has significantly increased and as a result an individual is more susceptible to receive a higher exposure over the course of his or her lifetime. Pregnant female diagnostic studies have followed this trend as there has been a 25% increase in C o mputed Tomography (CT) imaging studies from 1997 to 2006.1 An increase in the number studies coupled with newer scanner technology, which has shown to maintain if not raise radiation exposure for a single exam, have resulted in significant increase in individual and collective effective dose.2 3 Most pregnant women that undergo CT exams will incu r a fetal absorbed dose of less than 50 mGy over the course of their pregnancy. This level of absorbed dose is the recognized threshold for concern, as it is less than what is needed to induce abnormalities such as growth and mental retardation, which are common effects documented in the epidemiological studies of the Japanese bomb survivors. However, this level of exposure is comparable to radiation workers who by law require individual monitoring. For an occupational worker who is declared pregnant, the limit is 5mSv per gestation or 0.5mSv in any one month of

PAGE 15

15 pregnancy .4 Given the sensitivity of young children to radiation compared to adults, this discrepancy in monitoring needs to be addressed in some form. Using anatomically accurate reference physical models and recently developed advanced dosimetry techniques, the absorbed dose to the fetus and embryo will be evaluated at multiple gestation stages from clinical CT studies performed at the conventional imag ing protocols using a modern scanner The results of these studies are compared to both current biological data to validate or disprove claims about radiation safety in diagnostic radiology, and current clinical fetal/embryonic dosimetry methods. Most of these resolved uncertainties stem from the effectiveness of tube cur r ent modulation imaging and the accuracy of dose calculations using scanner reported DLP values. The dissertation is also intended to advocate any additional imaging precautions, or neces sitate new and advanced techniques of monitoring conceptus doses from clinical CT studies.

PAGE 16

16 CHAPTER 1 SUMMARY OF ESTIMATING CONCEPTUS DOSE, EVALUATING STOCHASTIC RISK, AND THE EFFECTS OF RADIATION ON UNBORN CHILDREN Clin i cal Dosimetry and Risk Unshielded i rradiation delivered to an unborn child in diagnostic radiology is avoided completely or limited to situations where the mothers health is severely jeopardized, as evidenced by the practice in clinical medicine of ALARA (As Low As Reasonably A chievable). The concept of ALARA is especially appli cable to C omputed Tomography (CT) studies which cumulatively account for a significant portion of the populations annual effective dose received in diagnostic studies.2 5 In order to minimize the fetal dose, clinical protocols are adjusted and additional precautions are utilized. Some of the modified imaging techniques include: using a filter that adds hardening to the beam, collimating the x ray fan beam to the diameter of the patient, collimating the z axis beam width to prevent wasted radiation, reducing the timecurrent product directly minimizing the s can length to o nly including critical anatomy, and the use of tubecurrent modulation scanning techniques .6 The extent to which these safeguards reduce the fetal dose in the later stages of pregnancy is based on the estimated conceptus dose studies in the literature. Before examining the risk of fetal doses from imaging studies, it is worthwhile to look at the recommended dose limitations that are currently established for pregnant females who may be potentially exposed to radiation. Currently the National Co uncil on Radiation Protection and Measurements ( NCRP ) recommends that the effective dose to the embryo or fetus of an occupational worker not exceed 0.5 mSv per month of gestation.3 The International Commission on Radiat ion Protection (ICRP)7 claims that no more than 2 mSv of equivalent dose should be delivered to the surface of the

PAGE 17

17 pregnant womans abdomen during her pregnancy In the case of diagnostic imaging, where low k Vp x rays are being used the se quantities can also be represented as 0.5 and 2 mGy of absorbed dose, respectively. It should be noted that in x ray dosimetry the equivalent dose (in m Sv ) and absorbed dose (in m Gy ) are numerically equivalent as the radiation weighting factor wR, for x rays and electrons is equal to unity Although these regulations are enforced for pregnant occupational workers, they do not apply to medical imaging. The benefit of a diagnosis or treatment that utilizes radiation will outweigh the risk from receivi ng radiation, especially if that risk is stochastic in nature. Clinical CT D osimetry The helical nature of the x ray source in multislice computed tomography requires a unique metric to evaluate x ray exposures from a specific clinical CT study Application specific d ose metrics have been introduced as a measurement that can be easily performed that accesses x ray tube output and generally correlates with dose for the specific imaging modality Since the 1970s Computed Tomography Dose Index (CT DI) has been the standard metric for quantifying the x ray tube output from a scanner operating under a specific imaging technique. The CTDI is defined as the average measured air kerma to the center of a cylindrical acrylic phantom For a multislice CT scanner capable of simultaneously scanning ( N ) of nominal slice thickness (T), the measured air kerma from a single full rotation along a line parallel to the axis of rotation, over a length of 100 mm, can be theoretically expressed as : = 1 ( ) ( 1 1 ) The instrument ation to measure the integral of the single rotation dose profile, D(z) is a 100mm ion pencil chamber which measures an integrated exposure reading

PAGE 18

18 over its active length (100mm) The integral in Equation 11 simplif ies to the measured exposure by the chamber corrected for air kerma using the f factor multiplied by the chambers active length. There are two types of CTDI phantoms, a head (16cm diameter ), and body (32cm diameter ) phantom. Both phantoms are acrylic c ylinders that have measurement points for the ion chamber on the periphery of the phantom at orthogonal angles and a single location at the center. For helical scanning, the definition of CTDI was modified to account for beam overlap or beam spacing, by a ccounting for the exam pitch. Measurements are taken at both the center and periphery, and weighted accordingly, and adjusted based on the pitch, to find the CTDI volume ( CTDIvol) CTDIvol represents the average dose across the area of the central scan plane (z=0). It is defined in terms of in Equation 12. = 1 3 + 2 ( 1 2 ) The calculation of takes into account mu ltiple locations within the phantom plane vs. a single axial measurement However it is really only a truly accurate measure if the pencil chamber can capture the primary beam and the scatter tails. Therefore with regards to wide beams the accuracy of this method becomes uncertain. To provide a more realistic quantity of the total radiation exposure, the Dose Length Product (DLP) was introduced to account for the scanned length. Thus for a multislice helical scanner, with a scan length ( l ), the DLP is defined as : = ( 1 3 ) The scanner reports the DLP value based on the beam profile and output tables generated by the manufacturer, and the specified technique for that exam: kV, mA,

PAGE 19

19 pitch, rotation time, and scan length. The values for DLP and CTDIvol are estimated values and not intended to explicitly represent any kind of patient dose. Therefore, it is common practice to measure CTDIvol and DLP for different scan techniques. These measurements are not used to calibrate the scanner, but merely to provide a secondary measurement of radiation output These values are very crude estimates, and are already outdated since there is no way to accurately measure CTDIvol for a 16 cm wide beam (320 slice scanner ).8 The CTDI methodology in quantifying radiation exposure is unique to CT. Other dos e quantities include entrance skin air kerma (ESK) in radiography and the kerma air product (KAP) in fluoroscopy. Both of these quantities are measured free in air using a designated ion chamber. All of them are independent of each other and require much further corrections to obtain any type of patient specific absorbed dose. One way to compare patient doses for various imaging modalities is to utilize the effective dose (ED). Therefore, the DLP to effective dose coefficient (ED/DLP) was developed to evaluate CT patient doses with respect to other imaging modalities and to provide organ dose data to evaluate stochastic risk. In a journal publication by Huda9, DLP values for different scanners and effective doses calculated by Monte Carlo dosimetry codes were implemented into a software program to estimate effective dose. Table 11 shows a comparison of three available dosimetry software programs that uses these parameters to estimate the effective dose from CT scans performed frequently in the clinic .10 All three software packages used the mathematically constructed stylized phantoms These computational phantoms are different from any real patient since they are based on standard height, weight, organ size and location. Also, the m ean effective dose

PAGE 20

20 and DLP are given for a reference sized individual weighing approximately 70 kg. The study acknowledges that when scanning patients whose size differs from 70 kg, different correction factors will have to be used if reliable ED values are to be obtained. The ED/DLP conversion factors overall provide a relative risk quantity but do not give an absolute risk assessment, lending itself as a useful method for determining the risk from exam to exam. There are many limitations when using the CTDI methodology for patient dosimetry especially when applied to newer generation scanners Most obvious, is the geometry of a typical patient is not very similar to a 32cm diameter cylinder T he dose profile within a slice is highly dependent on the diameter of tissue within that slice. This was explained in detail by a series of studies correlating radiation absorbed dose in CT to effective diameter. As previously stated, CTDI cannot be quantified for volumetric scanners which utilize beams collimated to widths (FWHM) greater than 100 mm. The third limitation is that CTDI underestimates dose. Most clinical scans lengths are greater than 100mm, and the scatter contribution from the regions outside the volume occupied by the chamber and phantom will contribute to the accumulated dose. At scan lengths greater than 25 cm, the dose approaches a limiting equilibrium. CTDI100 underestimates the equilibrium dose by a factor of 0.6 on the central axis and 0.8 on the periphery.8 11 CTDIvol places a higher emphasis on the periphery doses which may help offset the underestimated equilibrium dose since most organs lie in the middle of the patient When it relates to exams that utilize tube current modulation the quantit y CTDIvo l becomes more of an issue. The nature of CTDIvol is the average dose to a phantom at the central scan axis (z=0) A ll of the slices are assumed to be similar in

PAGE 21

21 geometry and the scan parameters remain constant throughout the entire scan, an accurate est imate of dose can be estimated The issue when quantifying CTDI in the traditional sense is that the spatial dose profile of the scan changes from the modulated tube current along the scanning axis. Boone and Dixon showed that three completely different T CM protocols can yield similar CTDIvol values.12 Th e dose profile cannot be captured by the ion chamber from a single rotation scan at a constant mA, therefore the value of CTDI canno t be accurately quantified from just a single measurement. Even if it could evaluated, the parameter would be spatially dependent and therefore a series of values would be required corresponding to a slice, and then matched with the target organ in order to correlate with absorbed radiation dose. Clinical Fluoroscopy Dosimetry As a means of off ering a comparison between CT and most of the other imaging modalities, a section on fluoroscopy was included. The fluoroscopy tube head contains a Roentgenarea product (RAP) meter mounted below the collimator to measure the intensity of the x ray beam striking the patient. The RAP meter is sensitive to exposure rate, time the machine is imaging the patient (fluoroscopy time), and area of the beam. Other imaging parameters such as kVp, magnification mode of the image intensifier, and sourceto patient distance are typically not recorded by the RAP meter. Skin dose is the one of the main concerns in fluoroscopy due to long exposure times, leading to significant r eported skin doses. The RAP meter is a useful tool for estimating the skin dose as the dosearea product can be mapped onto the surface of the patient. Applying correction factors, the skin dose may be calculated. The x ray source in fluoroscopy is deli vers radiation in projections relative to the patient The conceptus dose is calculated for a single projection using a manual

PAGE 22

22 technique facilitated by using T he Book Handbook of Selected Tissue Doses for Projections Common in Diagnostic Radiology.13 The output of the x ray tube for a given tube potential and filtration, or HVL, is measured by an ion chamber 100 cm f rom the x ray source. The parameter is measured in ionization (mR) per timecurrent product (mAs). The entrance skin exposure (ESE) for each projection can then be calculated using Equation 14. = 100 ( 1 4 ) W here the source to surface distance (SSD) and mAs are measured or given for each projection. The handbook lists tissue dose per ESE at different projections for different organ sites tabulated with various beam HVLs. While this method accounts for many parameters, patient characteristics such as size, shape, and body composition are ignored along with the collimated field size. These imaging parameters significantly contribute to the back s cattered radiation which is adds additional exposure. General R isk Assessment The effective dose (ED) is a dose quantity related to evaluating stochastic effects. It is the standard metric when estimating cancer risk from a nonuniform radiation exposure. The effective dose is calculated by totaling the tissuewei ghted organ doses of specific organs in the body that are most susceptible to stochastic effects from ionizing radiation. The formula, originally defined in ICRP 607, for calculating effective dose from x rays is shown in Equation 15 = ( 1 5 )

PAGE 23

23 Where is the tissue weighting factor, and is the equivalent dose (expressed in mSv) derived from the average organ absorbed dose multiplied by a radiationweighting factor specific to the incident radiation energy and type. Photons and electrons are always assigned a radiation weighting factor o f 1 therefore can be expressed as the absorbed dose averaged over the entire organ. ICRP 10314 is the revised publication that incorporates updated the tissueweighting factors for all risk contributing organs These factors are a relative risk assessment of each organ with respect to the others. The breast tissue weighting factor for example is 0.12, but the brain is assig ned a value of 0.01. By definition it would take approximately twelve times the radiation to produce the sa me stochastic event in the brain, as it would for the breast. The tissue weighting factors are derived from a uniform whole body exposure. Weighting is based on individual organs relative contribution to the total detriment. According to the ICRP, the effective dose is defined independent of age and gender The definition of the effective dose lacks in specificity since the relative radiosensitivity of organs differs greatly with age. The case studies of the Chernobyl accident revealed a very high inc idence in thyroid cancer for children versus a moderate prevalence in adults. Other examples include the higher risk of leukemia for children and high incidence of lung cancer relative to breast cancer for women over 60 versus that of an infant. The hig h contribution of the breast to the total detriment significantly impacts the effective dose calculation. The breast weighti ng factor is significantly over estimated for men and underestimated for women. Additionally, there is a significant difference in the amount of breast tissue between a 5 year old and a 30 year

PAGE 24

24 old woman. Another issue or common criticism of the effective dose is that it incorporates physical and hereditary effects, nonfatal and fatal cancers, into one factor for each organ over the entire population. Each of th ese end points should correspond to a different weighting factor, and any attempt to apply a weighted average of these endpoints is both arbitrary and subjective.15 Given the apparent differences in both anatomy and development between a fetus and an aged adult and the partial irradiation characteristic of these diagnostic studies the effecti ve dose is not an effective metric for evaluating stochastic risk for a developing fetus from CT exposures. As a result most radiobiology studies report fetal radiation absorbed dose rather than effective dose. BIER VII The most comprehensive risk mana gement study is by the BEIR VII16 committee which published its most recent report in 2005. The report published relative risk coefficients for a variety of organ specific solid tumor sites, whole body, and leukemia relating low leve ls of ionizing radiation to cancer and genetic effects. These empirical valu es are age and gender dependent, and account for elapsed time after the exposure event The detriment is evaluated for three separate endpoints: incidence, fatality from radiation induced cancer, and severe hereditary effects. The term excessive relative risk corresponds to the additional risk contributed from a radiation exposure producing a radiation weighted absorbed dose, D The term absolute relative risk combines the rel ative risk to the baseline, or unexposed lifetime risk. The functional form of the doseresponse curve is described by Equation 16, where ERR denotes the excessive relative risk. = [ ] 60 ( 1 6 )

PAGE 25

25 The model is derived from the linear nothreshold (LNT) model. The term s is the sex dependent coefficient, D is the radiationweighted dose (solid tumor) or effective dose (leukemia) in Sv. The variable, e* is a parameter dependent of the age of exposure, while a is the attainted age, and lastly and are empirically tabulated factors. The BEIR VII report found that health effects from very low doses would be difficult to evaluate due to contributions from other spontaneous factors, which are also statistical in nature, interfering with extrapolation of the data to these low dose levels. This noise threshold described in the report is stated to be 100 mSv, although the risk is expected to follow a linear trend with no threshold. Therefore any measured dose under 100 mSv would impose statistical limitations on identifying the root cause, but the risk dose relationship will presumably remain linear. Studies Evaluating Conceptus D ose A study by Jaffe17, was conducted to correlate the relationship of absorbed fetal dose in the early first trimester with CTDI vol when performing tube current modulation scans with variable noise index settings on a physical anthropomorphic reference female phantom. M easurements were obtained using an anthropomorphic female phantom, and segmenting the uterine wall to function as the dose point. Dose was measured using a metal oxide semi conductor field effect transit ( MOSFET ) detector, along with the reported scanner values for CTDIvol. Low dose protocols at 14 0 kVp were parameterized for a 16 slice m ultidetector CT (MDCT) GE scan ner (Lightspeed 16, GE Healthcare, Madison, WI) The results showed a linear correlation of CTDIvol with absorbed fetal dose. The reported absorbed dose ranged from 9.25 to 37.7 mGy. Studies by McCollough18 o btained similar measurements, also claiming conceptus dose

PAGE 26

26 from single acquisition CT does not in normal circumst ances exceed a nominal value of 35 mGy as seen in Table 2. Although Jaffe did evaluate the effects of TCM on a first trimester pregnant female, the modulation of the beam under these circumstances is minimal given the relatively uniform effective diameter of the abdomen and pelvic regions across the phantom Moving from a first trimester pregnant female to a second trimester should induce a significant modulation in the scanner output which could more accurately demonstrate the use of CTDIvol in TCM studies. Another study by Hurtwitz et al.19 evaluated the radi ation dose to the fetus at 0 and 3 months of pregnancy from a MDCT scanner. The radiation source was a 16slice MDCT scanner operating at standard protocols. Measurements were taken using a commercially available anthropomorphic female phantom modified t o represent a newly pregnant patient and a separate phantom that was 3 mo n ths pregnant. The physical phantom was used with a MOSFET dosimeter The reported fetal doses were as follows: renal stone protocol, 8 12 and 4 7 m Gy; appendix protocol, 15. 2 16.88 and 20 4 0 m Gy; and pulmonary embolus protocol, 2.4 4.7 and 0.6 0.6 6 m Gy. A study by Damilakis et al.20 used the Monte Carlo N Particle (MCNP) radiation transport code to evaluate the CT conceptus dose from two clinical scanners, Siemens Sensation 16 and Sensation 64 MDCT. Four mathematical phantoms simulated women at 0, 3, 6, and 9 months of gestation. Measurements were separately verified using an anthropomorphic phantom with a set of Lucite rings and TLD dosimeter s. The data was used to formulate a general for mulism for calculating the conceptus dose from the free air CTDI value, abdomen perimeter, and conceptus distance. The most comprehensive study, by Angel et al.21, generated 24 patient specific voxelized models

PAGE 27

27 from a cohort of 31 pregnant patients, spanning a range of gestational ages who m previously underwent clinical CT abdomen and pelvic exams. This study account ed for the diverse position and location of the fetus and discovered a very important correlation. Fetal dose in CT correlated to both the maternal perimeter and centroid of the fetuss depth within the mother. Dosimetry was however limited to studies performing single acquisition abdomen/pelvis exams. Radiobiology In addition to stochastic effects, developmental effects are of great concern when assessing damage to an irradiated fetus or embryo. Developmental effects have a proven dose threshold unlike carcinogenesis which is probabilistic in nature. Animal studies, epidemiological studies from the Japanese survivors, and cohort studies pertaining to medical exposures have provided information on both stochastic and developmental effects from ionizing radiation. From this information developmental effects have been categorized into three categories: lethal effects, malformations, and growth disturbances without malformations. Lethal effects are induced by radiation before or immediately after implantation of the embryo into the uterine wall, and increasingly higher doses thereafter. Embryo or fetal death can be categorized as either prenatal (before birth) or neonat al (at birth). Malformations are characteristic of interference or disturbances of organ development. Growth disturbances can occur at any time over the course of the pregnancy, but malformations are characteristic of radiation damage during early stages of gestation.

PAGE 28

28 Gestational Periods In order to better categorize risk from radiation, Russell and Russell divided the total developmental period into three stages: preimplantation, organogenesis, and fetal period.22 Preimplantation Preimplantation with regard to associative risk from radiation exposure is best characterized as all or nothing, inferring a high risk of prenatal death, and a very low risk of any abnormali ties Th is period extends from the time of fertilization to the implantation of the embryo i nto the wall of the uterus. This period of time usually ranges from 8 to 10 days, but is ultimately defined by the zygote transforming into a blastocyst. The embryo is highly sensitive to radiation because most cells are simulta neously entering mitosis and there is a small existing cell population. Most radiobiological data from this period comes from animal studies. Since irradiation duri ng this period is all or nothing and death is prenatal it often goes undetectable or considered a failure to implant and therefore difficult to make to determine a root cause. Animal studies collectively have placed an overall LD50 at 1Gy.23 This value ranges considerably, (1.5 to 0.3 Gy) due to gene expression and effects of radiosensitivity caused by the cell cycle. T he human zygote is eng aged in mitotic activity when it di vides from a single cell into 2 cells. Mitotic activity, which is considered to be the most sensitive stage of the cell cycle, occurs 2 to 6 hours post conception. Post fertilization, the zygote is least sensitive to radiation because gene expression has yet to manifest itself. It is commonly accepted that malformations cannot be induced by radiation during this time period. Animal s tudies show that malformations exist, but do not accurately demonstrate a dose threshold.24

PAGE 29

29 Org anogenesis O rganogenesis which is the second phase of gestation, is the period during which most organs are actively developing It is during this period that congenital abnormalities associated with structural malformations are most evident. This time frame begins when the embryo has been implanted in the uterine wall and lasts until the 6th week post conception. Although a majority of deaths are neonatal, prenatal death is prevalent in early organogenesis. In early organogenesis, c ells are at their m ost differentiating stage, as the embryo is rapidly increasing in size. Embryos exposed to radiation during early organogenesis show the greatest intrauterine growth retardation. However, growth retardation is recoverable, both in size and weight, before or after birth.25 Fetogenesis The last phase of gestation is the fetal period. During this period structures are already formed and are actively growing in size. Common effects include damage to the hematopoietic system, liver, and kidneys. The gonads are also vulnerable to irradia t ion during this time period. The e arly fetus exhibits the largest degree of permanent growth retardation.26 The fetal period is charac terized by of the transformation of an embryo into a fetus. Dekban27 conducted a survey that looked at instances of pelvic x ray irradiation of pregnant women and concluded that irradiation after 30 weeks of gestation is not likely to produce gross structural abnormalities leading to a permanent handicap in early life, but functional disabiliti es were possible. Irradiation of the fetus between 16 and 25 weeks of gestation may lead to a mild degree of microcephaly, mental retardation, and stunting of growth. Irradiation between 4 to 11 weeks leads to severe abnormalities of many organs in most children, most frequently

PAGE 30

30 stunted growth, microcephaly, and mental retardation. Abnormalities associating with the eye, skeleton, and gonads are also frequent, but in literature are not as frequent as the listed neurological disorders. Specific Developm ental and Carcinogenetic Effects Most malformations induced by low levels of radiation are cephalic, while other types are deemed limited. This is the result of migration of brain cells within the Central Nervous System (CNS) during the early development of the CNS. Data from the Japanese bomb survivors reveal microcephaly, mental retardation, and growth retardation as the primary effects.28 Radiation related mental retardation is limited to exposure during 825 weeks post conception, while most pronounced between 815 weeks. This time frame represents a period in which there is greatest proliferation of immature neurons and their migration from the ventricular proliferative zones to the cerebral cortex. The Japanese survivor data also demonstrates a dose threshold, which studies have determined to fall between 100 to 300 m Gy.29 Note that these valu es for absorbed embryonic/fetal dose are estimated. Another study concluded an IQ decline of 21 points/Gy for the 8 to 15 week period, and 13 points/Gy for the 1625 week period.30 ICRP7 concluded that radiationinduced mental decrement is det erministic with a threshold related to the minimum shift in IQ that can be measured. Using the previous data and fitting it to a linear threshold model, the threshold would result in about 50 m Gy, which hasnt been validated. Existing data has provided no clear evidence for a dose threshold with respect to IQ loss. Microcephaly, decreased head diameter, is observable only for exposures occurring in the time frame 1 to 16 weeks post fertilization, with no significant excess in risk seen after 16 weeks. This morphological finding was most prevalent in the

PAGE 31

31 Japanese bomb survivors data. Microcephaly is not prevalent during the 1625 week period when mental retardation can occur. As is the case, mental retardation doesnt appear visible in the 18 week time period. T he highest risk of mental retardation occurs when the cerebral cortex is being formed, which occurs after 8 weeks Cells killed before this period can cause small head size without affecting neuron development because the neurons that lead to t he formation of the cerebrum are not yet sensitive to radiation, unlike the glial cells which provide structure for the brain which tend to be susceptible for depletion. Microcephaly is a specific growth defect, and is related to the number of damaged or killed cells. No current evidence supports a precise dose threshold.31 BEIR VII claims that the minimum dose at which this effect has been observed in humans falls between 100 to 190 mGy. Carcinogenetic risk from inutero irradiation has been well documented and generally accepted as a liner nothreshold response. Human epidemiological studies have shown that inutero exposure to diagnostic x rays is associated with increased risk o f leukemia and, to a lesser extent, solid tumors.32 The root cause of solid tumors is difficult to ascertain due to the extended latent period of the cancer. These tumors are usually visible during the period of highest incidence which co incides in the later stages of human life further adding to the difficulty in distinguishing the mutagenic agent between ionizing radiation and a genetic predisposition. Gestational age has been shown to play a large role when accessing stochastic risk f rom prenatal irradiation. Animal data has shown a relative increase in risk during the early stages of gestation.33 This could be the result of scientific evidence that associates replication errors caused by radiation in proliferating cells. The embryo/fetus undergoes multiple cell divisions

PAGE 32

32 during its l ifetime, making it more sensitive to neoplasms. Similarly, the incidence of cancer is dose rate and LET dependent, which has also been well documented.31 One study, by the Childrens Cancer Group, found no evidence for an overall increase in risk for diagnostic inutero radiation.34 The authors credit this finding to a decrease in diagnostic x ray imaging involving pregnant females over the past decade. The OSCC (Oxford Survey of Childhood Cancer s) study35, which is the most comprehensive diagnost ic radiation study, calculated a cancer risk estimate of 640 excess cancers per 10,000 people per Gy and a 95% confidence interval between 4101000 excess cancers. However, the Japanese bomb survivors have shown a lower excess cancer incidence probability The Japanese bomb survivors data was unable to incorporate leukemia into its findings, due to the early onset of leukemia. It is also contested by Wakeford and Little36 who suggest that the number from the OSCC study should be three times lower, due to artifacts in its most recent years. Cohort studies have shown that there is almost no risk to prenatal diagnostic radiation exposure. Although these studies are very precise experimentally, they do not represent a large population of childhood cancer cases and hence have very broad confidence intervals. ICRP 90 Co nclusions The ICRP 9025 publication is a docum ent that reviewed experimental animal data on the inutero effects of medical radiation exposures, and evaluated a series of human studies concerning the inutero risk of cancer incidence and effects on the developing brain. Regarding carcinogenesis and m alformations, the following conclusions were drawn for a prenatal irradiation below 100 mGy of low LET radiation: In retrospect of the induction of malformations, there are gestational agedependent patterns of inutero radiosensitivity with maximum sensit ivity being expressed during the period of major organogenesis, and on the basis of animal

PAGE 33

33 data, there exists a true dose threshold of around 100 mGy for the induction of malformations. With regard to neurological effects, the A bomb survivor data stro ngly supports a dose threshold of 300 mGy for severe mental retardation during the most sensitive prenatal period (815 weeks post conception). The associated data on IQ loss is difficult to predict, and a nonthreshold dose response cannot be excluded, h owever IQ loss following the exposure of under 100 mGy would be of no practical significance. Regarding cancer risk, it is acceptable to correlate lifetime cancer risk following in utero exposures to irradiation in early childhood. The National Cancer Institute developed a risk assessment calculator based on the risk models proposed by 2006 BEIR VII publication with additional risk models developed by NCI for eight cancer sites, as well as additional cohort studies to update the thyroid and breast models.37, 38 The calculator employs a set of techniques to evaluate the estimated lifetime risk from a distribution of potential lifetime risks taking into account statistical uncertainties associated with the risk model parameters and subjective uncert ainties due to a number of assumptions. Table 13 lists a series of organs and tumor sites along with the calculated mean relative risk of a 0 to 1 year old male or female exposure to a uniform 50 mGy radiation absorbed dose. From Table 13 the most prevalent types of cancer are leukemia, and cancer from the liver and thyroid. The last two organs are weighted individually according to ICRP 103 as 0.04 of the total determent. In addition to those organs and the stomach, the other sites are equal at producing a carcinogenetic event. With this mind, it further demonstrates the lack of specificity in using the effective dose with regard to the fetus. The term relative risk is somewhat misleading given that the baseline risk of childhood cancers is extremely low and the lifetime risk of cancer over the course of ones lifetime is very high. Therefore, e stimating cancer risk is very cu mbersome due to

PAGE 34

34 the difficulty in ascertaining the cause of cancer from radiation or genetics due to similar attained ages o f carcinogenetic expression which is occurs mostly in the later part of life In addition the epidemiological and cohort studies that use these models show wide confidence margins and apply a small statistical sampling to a large population as shown by th e vast confidence intervals in determining ERR from the last column in Table 13 Although not listed The Health Physics Society39 states that with regard to risk assessme n t : Radiological risk assessment, particularly for radiogenic cancer, currently is only able to demonstrate a consistently elevated risk in the intermediateand high dose groups of the studied populations Cancer and other health effects have not been observed consistently at low doses (< 100 m Gy), much less at the even lower doses (< 10 m Gy) typical of most occupational and environmental exposures. Consequently, in order to estimate radiation risk in the low dose region, observed health effects in the higher dose regions are extrapolated to the low dose region by using a variety of mathematical models, including the linear, nothreshold model (with a correction for dose and dose rate ). With regard to risk assessment the radiobiology is completely different between a developing fetus and adult, and therefore what is applicable to one may not necessary to be applicable to other. A fetus and embryo are physiologically undergoing more cell differentiation, migration, and proliferation than an adult, making it more susceptible to low levels of ionizing radiation and therefore posing a higher ris k.40 Summary of Effects Developmental effects induced by radiation exposure to a developing embryo in CT are possible. The two most prevailing and common effects documented in literature are mental and growth retardation. The minimum dose threshold for mental retardation is stated to fluctuate between 100 and 200 m Gy. This taking place between the 816th week gestational period. Any exposure under 100 mGy falls below the established

PAGE 35

35 threshold for the formation of any congenital abnormalities. With regards to deleterious effects, the American College of Obstetricians and Gynecologists41 published the following opinion: Women should be counseled that x ray exposure from a single diagnostic procedure does not result in harmful fetal effects. Specifically, exposure to less than 5 rad [50 mGy] has not been associated with an increase in fetal abnormalities or pregnancy loss. A report conducted by the Health Protection Agency of the United Kingdom found that t he natural childhood cancer in the UK is reported as 1 in 500, and a fetal exposure as small as 10 mGy was claimed to double the baseline risk .42 From Table 84 in I CRP 84, 50 mGy would result in no increased risk of abnormalities and an ERR of 3.0 %.43 Sufficient biological evidence is not available to support a claim that a single CT examination, or a conceptus dose under 50 mGy, would jeopardize the development of a fetus or embryo. Finally, the effective dose is a completely ina dequate method in evaluating low dose stochastic effects, as it compounds the error in accessing risk to an unborn child from a diagnostic radiation exposure. It also overestimates and underestimates the detriment contribution from individual organs at low doses. From Table 13, the breast, colon, and lung contribute a lot less to the overall risk than the thyroid and liver. The 50 mGy benchmark has been well documented to exceed the fetal dose received from most CT examinations, on the account of the vast maj ority of routine single phase acquisition computed tomography studies deliver less than 35 mGy to the uterus. To achieve a conceptus dose greater than 50 mGy, a pregnant patient would have to undergo multiple CT exams. Practically, no medical clinician w ould allow this to happen if he or she had knowledge the patient was pregnant, leading to the scenario

PAGE 36

36 where the patient and doctor are unaware of pregnancy, and the patient is subject to several examinations. The rate at which pregnant women are unintent ionally exposed to radiation is largely unknown. One study involving trauma patients concluded that 2.9% of the women between the ages of 15 and 40 years admitted to the trauma unit were found to be pregnant. Of the pregnant women who were admitted, 11% were incidental pregnancies, of which 8% were newly diagnosed.44 Research Objective For the purposes of patient dosimetry risk it can be assumed that any exposure resulting in less than 5 mGy, the occupational limit for the declared pregnant worker, can be assumed to be acceptable since this limit is determined without accounting for the clinical benefit of receiving a diagnostic test to the overall welfare of the mother. This manuscript is not intended to justify alternative imaging methods or promote a dose limitation. The research in this study simply attempts to ad vocate the need to monitor radiation exposures from clinical CT for females that are or may be potentially pregnant, in the same way we monitor radiation workers. In addition, this research will examine the accuracy of current methods of estimating doses with the intent on validating their place within the clinic. The information presented in this chapter highlights the significance of monitoring and evaluating embryonic/fetal doses in CT The overall objective of this dissertation is to advance the phys ical tools and methods used currently to quantify conceptus doses associated with low levels of ionizing radiation used in current CT procedures than the studies presented in this work The results will be used to provide a framework that will allow medical personnel to conduct clinical CT studies so that they may result in minimal risk to the women and unborn child. Each specific aim is highlighted below:

PAGE 37

37 1. Construct a fiber optic dosimeter that is independent to v ariation in depth and energy and can easily be integrated with anthropomorp hic phantoms. In addition, the dosimeter is able to measur e absorbed dose to a variety of bodily tissues. C haracterize the dosimeter to verify that it maintains a dose rate and angular independence, and high sensitivity and reproducibility to radiation seen with water equivalent FOC dosimeter s. 2. Characterize the overall error to using anthropomorphic phantom materials for dosimetry studies, and computationally evaluate the sensiti vity of measuring dose with these materials in comparison to their tissue counterparts. This goal includes the develop ment of a new methodology to expedit e the phantom construction process, eliminate the intermittent anatomical structures that results fro m building the computational phantom from axial slices, and create dose points within the phantom in the sagittal and coronal anatomical planes allowing for dose points in all three anatomical planes. 3. Develop a dosimetry methodology for measuring absorbed dose to anthropomorphic phantoms that incorporat es the linear fiber optic dosimeter and the energy independent dosimeter. The linear fiber optic dosimeter is a device that has been characterized to exhibit minimal variation in sensitivity to ionizing radi a tion along its axis, making the ideal dosimeter for measuring dose in one dimension. By combining multiple s the dose across the entire volume may be evaluated more accurately. 4. F etal/embryo dose measurements, using each pregnant female phantom and the dosimeter s discussed above performed using a clinical CT scanner. The scan protocols administered were selected based on their clinical significance and the level of radiation delivered to the abdomen and pelvis. 5. Use collected data to benchmark modern m ethods of estimating the conceptus dose received from clinical CT studies examined in aim 4. The following chapters of this dissertation are dedicated to addressing each of the aims discussed above in greater detail. Each chapter corresponds to a specif ic numbered project above and gives the pertinent background information, experimental methodology, and results. An extra chapter was added to summarize the effects of these results and how they pertain to the clinical use of CT on pregnant and potentiall y pregnant females. Another chapter was incorporated to address the issues of using fiber optic dosimeter s in high energy applications specifically external beam photon and electron radiotherapy.

PAGE 38

38 Table 11. Comparison of effective dose calculated from C T dose software for female from multi detector CT (M D CT). Protocols selected based on European guidelines .10 Dosimetry Effective Dose (mSv ) Software Head Chest Abdominal Pelvic WinDose 1 8 8.8 12.1 10. 5 ImpactDose 2.2 11.0 13.0 14.0 CT Expo 2. 6 13.8 16. 4 16.5 Coefficient of Variation (% ) 17.5 22.1 16.4 22.2 Table 12. Conceptus d oses from a sample of common clinical CT studies B ottom three protocols indicated nominal doses exceeding 2 mGy.18 Examination Fetal Dose Protocol [mGy] Head CT 0 Chest CT 0.2 CTA (chest) 0.1 Routine Abdominal 4 Routine Abdomen/Pelvis 25 CTA (chest thru pelvis) 34

PAGE 39

39 Table 1 3. Lifetime risk from a 50 mGy acute exposure to a 0 1 year old child. Organ/Site Gender Excess # of cancers Baseline # of cancers Total # of cancers ERR (confidence interval) All Organs M 1990 48290 50280 3.96% (2.00% 6.30% All Organs F 3350 41230 44580 7.51% (4.33% 12.32%) Brain/CNS M 69.7 661 730 9.55% (2.25% 23.68%) Brain/CNS F 19.6 537 557 3.52% (0.93% 8.29%) Breast F 819 13176 13994 5.85% (3.17% 8.51%) Colon M 242 4079 4321 5.60% (2.91% 9.25%) Colon F 165 4200 4364 3.78% (1.62% 7.21%) Leukemia M 177 939 1116 15.86% (1.65% 57.30%) Leukemia F 155 722 876 17.69% (1.22% 69.01%) Liver M 77.6 833 910 8.53% (1.35% 29.90%) Liver F 43.3 349 392 11.05% (0.83% 44.54%) Lung M 234 7931 8165 2.87% (1.25% 5.28%) Lung F 552 6326 6878 8.03% (3.81% 15.64%) Mouth M 38 1427 1465 2.59% (0.29% 6.04%) Mouth F 36.5 683 719 5.08% (1.17% 11.19%) Ovaries F 74.5 1445 1519 4.90% (1.23% 12.17%) Prostate M 139 16385 16523 0.84% (0.00% 5.32%) Stomach M 121 1095 1217 9.94% (0.76% 39.59%) Stomach F 148 683 831 17.81% (1.59% 73.35%) Thyroid M 134 376 510 26.27% (5.01% 77.72%) Thyroid F 742 1070 1812 40.95% (7.08% 137.88%) Uterus F 46.3 3241 3287 1.41% (0.00% 4.71%) all cancers listed are mean number of cancers for a population of 100,000 people within a 90% uncertainty range ** acute 50mGy exposure to a person who is born in 2014 and exposed in 2014 *** confidence intervals calculated from gi ven upper and lower bounds for e xcess # of cancers and total #

PAGE 40

40 CHAPTER 2 HIGH ENERGY COUNT RATE EFFECTS AND BACKGROUND REMOVALISSUES IN PLASTIC SCINTILLATING FIBER OPTIC DOSIMETRY Counting Issues in Fiber Optic D osimeters Used in Rad i otherapy Dosimetry The stem effect is traditionally a term coined for the additional charge created outside the sensitive region of an ionization chamber or diode and subsequently measured contributing an excessive signal. I n most situations the ionization chamber s stem is usual ly in the radiation field, therefore a correction is not needed when the calibration is performed with the stem completely in the field. For fiber optic detectors this value is significantly dependent on the amount of optical cable in the radiation field, which is highly variable.45, 46 With respect to fiber optic s used in radiation detection, the stem effect refers to the natural fluorescence of optical fibers under exposure to ionizing radiation leading to additional signal contribution. A portion of the light generated within the optical fiber is collected adding to the signal from the scintillator Not properly accounting for the optical fibers light contribution can impact the accuracy of measurements especially under varying x ray field sizes.4749 Plastic scintillating dosimeter s (PSDs) have many charac teristics that prove to be very beneficial for radiotherapy dosimetry. Their sensitive volume, polystyrene, has been determined through Monte Carlo methods to be energy independent for x ray photons above 500 keV.50 In addition, they can be fabricated into miniature dosimeters making them very useful for internal in vivo dosimetry. Plastic scintillators are very sensitive to ionizing radiation and can therefore be made with little material to produce an adequate s igna l. Their main drawback is the optical fiber component whic h is an integral component of the PSD system, produces a prompt and substantial signal when

PAGE 41

41 exposed to ionizing radiation. This is signal is significant enough to severely affect the accuracy of radiation dose measurements, especially in the megavoltage energy range. Hyer examined the applicability of using PSDs for radiotherapy dosimetry.51 H is characterization of these detectors and the work of Beddar52, 53, have demonstrated excellent dose linearity and reproducibility as well as an a de quate response to varying pulse repetition rates However, the findings uncovered a dependence related to the radiation field length, when using the background subtraction method. This dependence hasnt been discussed in great detail or observed in other publications The literature has somewhat successfully demonstrated the use of PSDs under small field sizes for IMRT and Stereotactic radiosurgery modalities as well as the achievements of chromatic filtering using capable devices such as CCD camera s.5456 T his chapter will examine how the bac kground signal from an optical fiber a ffects the quality of dosimetry for a detection system using a plastic scintillator, photomultiplier tube array and no use of any chromatic filtering or spectral analysis techniques which recently have proven to be quite successful.47, 56, 57 The Stem Effect Cerenkov Radiation The stem effect is much more observable at megavoltage energies than at kilovoltage energies because of Cerenkov radiation. Cerenkov radiation is electromagnetic radiation emitted when an electron passes through a dielectric medium at a velocity greater than the speed travels in the medium. As a result of this phenomenon, more background counts are detected by the dosimetry system. The threshol d for Cerenkov radiation has been estimated to above 178 keV in PMMA and 144 keV in polystyrene.53, 58 In radiotherapy applications, using background subtraction

PAGE 42

42 as a Cerenkov removal technique has shown some success.53, 54 59 This method is ultimately limited by the spatial resolution of the two fibers under the presence of a strong dose gradient perpendicular to the fibers, as its ultimate success is dependent on equal irradiation of each optical fiber. The collection of light produced from Cherenkov radiation is highly dependent on the direction of the charged particle. In contrast to light produced from scintillators, in which light is emitted isotropically, Cherenkov photons are emitted in the direction of the particles path. The light is releas ed in a cone defined by an angle ( ) which is defined as : = 1 ( 2 1) The variable is the ratio in velocity of the particle to the speed of light, and n is the index of refraction of the medium. All light produced inside the optical fiber cannot be collected by the PMT because the mechanism for channeling light through a narrow dielectric (via total internal reflection) is limited to the angle defined b y the numerical aperture of the optical fiber. Light striking the walls of the fiber will be totally internally reflected, provided that the incident angle at each reflection is greater than the critical angle: = ( 2 2 ) Figure 2 1 illustrates the geometry of the optical fiber in relation to the incident ionizing particle and the totally internally reflected light propagating through the core of the optical fiber. The light collection cone of the optical fiber represented by can be expressed in radians as:

PAGE 43

43 = 2 ( 2 3 ) In most cases the optical fiber is positioned orthogonally to the direction of the incident radiation. High energy charged particles typically interact with matter in soft columbic interactions that do not affect the trajectory of the charged particle. Therefore, in the case of superficial megavoltage electrons and, more so for protons, the Cerenkov radiation generated in the optical fiber will lie outside the acceptance cone of the optical fiber. However in the case of x rays which interact mostly through Compton scattering, interactions will produce charged particles in the axially direction that fall under the acceptance cone of the optical fiber. Figure 2 1 presents t he trajectory of a charged particle inside an optical fiber along with the Cerenkov cone with respect to the acceptance cone of the optical fiber. For the Eska optical fiber ( Eska CK 20, Mitsubishi Rayon Co, Tokyo, Japan) the core and cladding index of r efraction is 1.49 and 1.403 respectively, which gives a cone with a vortex angle ( ) 0.303 radians, translating to a solid angle of 0.954 radians. Color Centers Irradiation of plastic scintillating fibers and optical fibers h as shown to affect t he ir optical properties.6062 The effect is a temporal and/or permanent loss in the transmission of light. The reduction in light transmission is attributed to the formation of color centers inducing an increase in optical absorption properties of wavelengths in the visible/UV light spectrum. A study examined the color formation by irradiation for a variety of polymers and found that all exhibited a shift in the UV cutoff and the formation of an absorption tail which can extend into the visibl e light range.63 The formation of color centers are characterized into two types: permanent and annealable, of which the

PAGE 44

44 latter is a recoverable process achieved by the release of trapped radicals via migration of free oxygen or radical radical reactions without the presence of air.64, 65 The annealing process through oxygen for polystyrene can be characteriz ed on the or der of hours up days post irradiation, and is sensitive to variation in dose rate.66 67 Any residual damage after this time period can be considered permanent. For polystyrene it was observed that transmission damage observed during irradiation is considerably larger than permanent damage especially for wavelengths in the visible light spectrum. The transient effects from recombination originate from the formation of electronhole pairs or free radicals generated by ionizing radiation within the lattice structure of the fiber.68 Most of these free radicals recombine; the remaining free radical species are trapped inside imperfections within the latt ice and may be released through other mechanisms, most notably oxygenation as discussed above. It is these trapped electrons and free radicals that affect the time dependent optical transmission properties of waveguides Following irradiation, electronh ole pairs are formed by ionizing radiation within the fiber initiating a transient absorption effect that results in immediate loss of light transmission. At lower doses of radiation the transmission of a fiber will increase towards the preirradiation v alue. The time between the transient absorption and the intrinsic absorption is known as recovery. The recovery kinetics and transient absorption was evaluated by Mattern et al.61 for four types of known radioresistant fibers: polystyrene, PMMA, Gedoped silica, and pure vitreous silica. The recovery kinetics for each fiber type follows a power law behav ior with around 90% recovery taking place within 1ms. The data was obtained using small radiation pulses at very high dose rates well exceeding those experienced clinically. However, the data

PAGE 45

45 strongly supported claims that the recovery kinetics of all the fibers examined after a transient absorption appear to be independent of dose, irradiation type, and wavelength which was evaluated between 500 to 800nm. It was also noteworthy that the peak transient absorption for PMMA and silica occurred at 400 nm and 550 nm respectively, and about 4 times lower in magnitude for the silica fiber thus making the PMMA core fiber more sensitive to optical absorption of Cerenkov photons. Ideal Stem Response to Radiation for a Pair of Optical Fibers The majority of the light reaching the PMT are photons released by the scintillator, while the remaining contributing portion is from the natural fluorescence of the optical fiber from either natural luminescence or Cerenkov mechanisms Luminescence is a result of deexcitation of orbital electrons or recombination of free radicals with electronhole pairs within the fiber. Alternatively the energy deposited may be in the form of heat. This additional amount of light may be accounted for by adding an optical fiber of equal length and placed parallel to the signal fiber This positioning would allow the fiber t o experience the same x ray intensity and energy spectra, and simplify the relationship between the stem counts in each fiber to a single factor. The number of detectable steam counts at any position along the optical fiber can be expressed fundamentally by Equation 2 4 ( ) = ( ) ( 2 4 ) w here N is equal to the number of counts detected from irradiation at position x. S is the scintillation efficiency of the glass fiber. I(x) is the x ray beam intensity along the fibers axis. A is the cross section area of the fiber and L is the length of the fiber being

PAGE 46

46 irradiated. The term is the fraction of light produced in the fiber that is detected by the PMT and can be further represented as a product of other variables = ( 2 5 ) Where: represents the fraction of light photons produced in the scintillator that successfully travel thru the optical fiber. This parameter is often associated with the acceptance angle characterized by the numerical aperture of the optical fiber. represents the transmission efficiency of light reaching the end of the optical fiber. Signal loss is attributed to light attenuation within the optical fiber. represents the detection efficiency of light photons detected by the PMT. Th is value is variable among other PMT devices and is dependent on the wavelength of the light photon. These additional terms represent the relative amount of light produced in the optical fiber that reaches the photon counter. For two optical fibers placed contiguous and parallel to each under uniform x ray field, the assumption of equal amounts of radiation is absorbed by each fiber at the same location is valid. Under a uniform radiation field, the relationship between the quantities of detected stem c ounts in each fiber can be expressed using Equation 2 6 = ( 2 6 ) Using Equation 2 6 the reference fiber can be calibrated to measure the stem counts in the scintillator fiber. The ratio of light collection efficiency may be obtai ned by placing the optical fiber inside a radiation field with the scintillator shielded and placed outside the primary beam while irradiating the optical fibers at varying energies and field sizes to determine the relationship in their responses. Equation 2 6 also describes an

PAGE 47

47 ideal response for the background subtraction method, which the measured counts in both fibers should correlate linearly with length of fiber irradiated. Count Rate Effects Radiation Delivery Mechanisms of Clinical Radiation Generat ors Continuous s ources Some clinical x ray sources are delivered from a beam current incident to a high Z medium target that produces b remsstrahlung radiation. These are typically radiological kilovoltage x ray tubes and orthovoltage tubes. Long lived radionuclides also generate a constant level radiation over a short time span. The dose rate is controlled by the current flowing into the target and the distance the absorber is from the high Z target. Under fixed conditions a detector placed within the x ray beam experiences a steady state source of constant intensity and produces a relatively stable count rate. L inear a ccelerator ( LINAC ) Bremsstrahlung radiation emitting from a LINAC is generated in pulses from the pulsing of the magnetron supplying the accelerator waveguide with oscillating microwaves for duration of several microseconds at a frequency of 3,000 MHz. A typical Varian Trilogy CLINAC calibrated to deliver 1 cGy per Monitor Unit (MU) at an SSD of 100 cm and a depth of 1.5 cm for a 6 MV beam produces an estimated instant aneous dose rate of 77.22 Gy/s (0.00028 Gy per pulse).69 This radiation source can be represented by the timedependent sketch in Figure 22 The larger pulses are separ a ted by the pulse repetition frequency (1/f) which is expressed in terms of monitor units per minute which clinically can fall between 100 and 1000 MU/min. Standard calibration protocols are set to deliver 100cGy per 100 monitor units at 100cm source-

PAGE 48

48 to detector distance (SSD) to an ionization chamber immersed inside a water phantom located at the depth below the surface corresponding to the maximum dose. Dead Time Effects f rom a PMT Detection System Potential c ounting losses may arise at high count rates from dead time behavior of scintillation counting systems Medical imaging devices such as SPECT and PET that use sensitive detectors with high absorption efficiencies require performance evaluation at high count rates for image distortion effects and loss of image contrast.70, 71 The energy to light conversion efficiency of polystyrene in the sub 100 keV range is relatively small compared to inorganic scintillators that are more sensitive to low energy photons due to their high atomic number and inherently higher photon yields. The smaller light yield of the plastic scintillator creates less dependence on the time resolution characteristics o f the photon detector, since the count rate is smaller dead time contributions are also smaller. The count rate linearity of a fiber optic system is determined by the resolution time of that system, which is defined as the minimum time between successiv e scintillation pulses that the detection system can distinguish. This time interval may be limited by the decay rate of the scintillator, the frequency characteristics of the signal processing circuit, and the timing characteristics of the photodetectors electronics. Photomultiplier tubes (PMTs) are photon counting devices that have an exceptionally fast time response.72, 73 The time response of a PMT is determined by the electron transit time. Electron transit time is the time interval between the arrival of the light photon at the cathode and its subsequent collection at the anode after being multiplied. The period between arrival and collection, however, only induces a fixed delay in the signal. The main concern with respect to dead time is the transit time spread (TTS).

PAGE 49

49 TTS is the temporal distribution of electrons arriving at the anode due to differences in electron pathlength and velocity. A wide distribution may lead to potential overlapping between pulses leading to loss of signal. The concept of TTS is illustrated in Figure 1. Transit time fluctuations in the electron transit time have two components: a chromatic dependence due to the spread of photoelectron initial velocities, and a geom etric dependence caused by variation in the distance traveled by the electron from the cathode to the anode. The chromatic dependence may be offset by operating the PMT at wavelengths corresponding to its peak sensitivity. This results from a larger numb er of photoelectrons being produced in the cathode, offering a better sampling of photoelectrons. Commercial PMTs used for photon counting require the use of additional electronic devices that process the output pulses from the PMT. Additional integrated electronic components may further modulate the shape of the individual pulses which may lead to pulse overlapping. Therefore, it is important to evaluate the timing properties of the entire counting system rather than individual components. Dead Time Modeling Counting losses incurred due to the time resolution may potentially be modeled using one of the two statistical classifications of dead time behavior: paralyzable and nonparalyzable.74, 75 It is widely recognized that scintillation camera systems have both nonparalyzable and paralyzable operating ranges. The nonparalyzable model predicts that the detector will be temporarily insensitive to radiation during the dead time, while in the paralyzable model, the detector remains sensitive to any additional events which may lead to pileup. In this scenario, two events separated by a time less than the dead time would result in neither event being detected. In the nonparalyzable case, the detector is able to detect the first event but not the second leading to saturation of

PAGE 50

50 detectable counts at high count rates. The true count rate may be analytically solved for by Equation 2 7 = 1 ( 2 7 ) I n Equation 2 7 m identifies the observed count rate, n the true count rate, and is the dead time. However, this model explains an ideal behavior as is subject to variation inevitably leading to inaccuracies, especially at high count losses given the sensitivity of n to changes in the product Since the system dead time is determined by the time resolution of the individual components of the system, any components that are not characterized as nonparalyzable may lead to a nonconventional model. Addition considerations need to be addressed when modeling dead time behavior from pulsed sources. Knoll76 addresses three cases where the dead time ( ) is less than T less than T but greater than 1/f, and greater than T but less than the off time. In all three cases the radiation intensity is assumed constant throughout the duration of the pulse. However none of these models may apply to our condition, which is complicated by the fact that two sets of pulses on two different magnitudes of frequency are occurring due to p ulsing from the magnetron and the electron gun. A model by Daams77 was derived for the apparent dead time of a detector measuring intensity of radiation which varies periodically = ( 1 2 ) ( 2 8 ) The Equation 28 is derived for a simple periodic function where and represent the average detected and true counting rate respectively In this case a maximum number of counts are observed which corresponds to The

PAGE 51

51 response of the detector to an oscillating source shows a linear increase in observed co unts as a function of frequency unit it reaches the maxima. Materials and Methods Detection System The plastic scintillating dosimetry system was constructed by Hyer and characterized for diagnostic energy dosimetry.78 The dosimeter consists of two cylindrical, 500 micron diameter multimode PMMA optical fibers (Eska CK 20, Mitsubishi Rayon Co, Tokyo, Japan) which were lengthened to 13 meters to account for LINAC vault rooms One of the fibers was coupled to a plastic scintillating fiber element, polystyrene. An uncoupled optical fiber is utilized to account for the uncontrollable radiation induced fluorescence and Cerenkov radiation discussed above. The scintillator was manufactured as a plastic cylindrical fiber (500 microns). Before coupling, the scintillator and coupling end of the optical fiber surfaces w ere polished using an Amphenol handpuck (Ocean Optics) and 3 grits of lapping paper ( Angstro m Lap, 12, 3, and 1 micrometer) to increase the number of light photons incident to the PMT (H7467, Hamamatsu Corporation, Bridgewater, NJ).79 The scintillator and optical fiber wer e manually coupled using a transparent adhesive gel (Crazy Glue) and between 1 and 2 cm of 3/64 PVC heat shrink tubing (Fisher Scientific) to ensure a tight coupling. After coupling, the plastic scintillator was cut to 2 mm from the coupling surface, an d polished. A reflector coating (Eljen Technology, EJ 510) was applied to the uncoupled end of the scintillator in order to increase the number of measurable scintillation photons.80 To provide structural and shield from ambient light, the two fibers were then encased in 1/16 heat shrink tubing.

PAGE 52

52 Considering the transient absorption effects as a possibility for apparent quenching of the detector from high energy r adiation, a gadolinium detector was constructed. Gadolinium oxysulfide (GOS) is a scintillator with a very long decay time, on the order of hundreds of microseconds, which would make the detector more sensitive by overcoming the initial absorption losses Using the same method as above, a of an image intensifying screen for clinical screen film mammography (Kodak MIN R 2 Cassette, Eastman Kodak Company). The GOS was coupled to a polished optical fiber by applying a thin coat of the adhesive gel to the polished end of the optical waveguide and placing the fiber onto the slab while exerting pressure to enhance the binding process. After a short time, the optical fiber was removed along with the attached thin layer of GOS coupled to the end of the fiber. A single PMT is required for monitoring each optical fiber. Therefore the fibers had to be separated from the heat shrink tubing to allow for connection with the PMT array. The uncoupled end of each optical fiber was mounted with a female SMA 510 connector (SMA 905) using a SMA epoxy glue (FiberFin, 1656 resin, #80 hardener). The remaining uncovered optical fiber was wrapped in heat shrink tubing. A male SMA optical fiber adapter (E577651, Hamamatsu Corporation, Bridgewater, NJ ) was used to interface the fiber to the PMT. The counting data was buffered and transferred through a hub to a computer for readout. A USB/serial hub (UPORT 16108, Moxa Inc, Brea, CA) provides an efficient communication pathway between the computer and the PMT array. Each PMT and the hub are powered by 12V and 5V DC power supplies, respectively.

PAGE 53

53 Sources of Radiation The following types of radiation sources were utilized to explain and offer a comparison in t he behavior of the plastic scintillation counting system at very high dose rates, as well as the ionization chamber used to accurately quantify the dose rate. A. Varian Trilogy CLINAC (2300IX, Varian Medical Systems, Palo Alto, CA) : 6 MV and 6 MeV photon and electron sources delivering radiation in pulse sequences described in Table I. B. Clinical Proton Accelerator at the University of Florida Proton Therapy Institute (Cyclotron, ESS, IBA, LouvainLaNueve, Belgium Delivers protons in a continuous beam current using a cyclotron with ESS between 25 to 300 nA. C. Cobalt 60 irradiator. Cobalt 60 source that can be controlled using a lever system to position the source at variable distances from the detector. D. Orthvoltage floor mounted unit (Gulmay Medical Limited, Suwanee, GA) with xray tube ( NDI 321, Varian Medical Systems, Palo Alto, CA ) selected imaging techniques between 150 and 320 kVp Sources labeled B C and D were used to characterize the dead time of the counting system us ing the nonparalyzable dead time model. Variable dose rates were achieved by altering the radiation current. In sourced B and C the radiation had enough energy to generate Cerenkov radiation, however a lesser extent in source B The true count rate was expressed as the measured dose rate which was evaluated using the appropriate ionization chamber. The same background correction factor was used for each experiment. It should also be noted that sources B and D were equipped to control the size of the radiation field which in turn could expose the optical fibers to varying known exposure levels of radiation. Determining the count rate from source A irradiations was difficult because measuring the count rate requires knowledge of the beam on time which isnt continuous and measurable with the available equipment. Therefore the dead time

PAGE 54

54 couldnt be measured directly but, could be tested to see if the nonparalyzable model would still hold true for pulsed sources. This was done by creating a modified version of the parameter to include the beam on time. The dose rate was varied by changing the position of the detector or using a water phantom to take measurements and provide different penetration thicknesses. However the number of pulses was kept relatively constant since the number MUs delivered at each position were unchanged. Experimental Methods Stem count behavior Equation 24 explains the ideal behavior of a pair of fiber optic cable s response to radiation. Under the presence of a uniform radiation field, the expected value of N would be proportional to the length of cable within the field. Sources A B and D are all equipped with c ollimators to define a field size measurable at a reference location. By p lacing the optical fiber within and parallel to the radiation field, measurements can be taken to evaluate I(x)dx. M ultiple readings may be acquired over a range of field sizes to determine the linearity of the background optical fiber to radiation. When considering the axial profiles of electron and proton megavoltage beams they are characterized as relatively flat, as well as diagnostic energy beams used in radiology. The spatial intensity profile of the radiation beam can be evaluated by using devices such as film and diode arrays. System t ime r esolution The time resolution, from sources continuously irradiating the dosimeter was evaluated by delivering a variable dose rate t o the scintillator, which can be achieved by controlling the beam current. To measure the true count rate, a dedicated ion chamber was used to quantif y the exposure rate and exposure time. The observed count rate

PAGE 55

55 and exposure rate were calculated by averaging the net scintillation counts or exposure and dividing by the exposure time. For this method, the true count rate is expressed in terms of the exposure rate, as the two are related by a conversion factor. Th e dead time was calculated by expressing the data in terms of the inverse observed count and the inverse exposure rate. Rearranging the terms in E quation 2 7 the dead time can be solved for using linear regression to find the optimal y intercept of the i nverse of the true count rate and observed count rate data: 1 = 1 n + (2 9 ) In E quation 2 9 the theoretical value of n is the product of the dose rate and a conversion factor relating the number of true detected counts to the measured dose. Dose and dose rate quantities may be evaluated with an ionization chamber specific for the treatment or imaging modality. In the megavoltage energy range, this value is independent of the energy. However, fortunately for this method, only the relative value of n is required to find the y intercept. An instrument such as ion chamber is suitable to find the relative true count rate. The error associated with using this fit can be calculated using Equation 35 which is discussed in more detail in Chapter 3. Gadolinium oxysulfide and the p lastic p erformance Each dosimeter was evaluated for PRF and field size dependence. The pulse rate dependence of each detector was investigated by delivering a constant dose at varying pulse repetition frequencies. Data was normalized to the lowest PRF. The field size dependence was investigated to find the effects of varying quantities of Cerenkov radiation by referencing the output factor. T he response of the optical fiber is proportional to the amount of radiation irradiating the optical fiber, which can be

PAGE 56

56 quantified in terms of the length of the optical fiber in the radiation field. Measurements were taken at five field sizes and normaliz ed to the 10x10 cm field size to represent the output factor for several widely used field sizes in radiation therapy. A 5 cm slab of solid water was placed below the detector to provide backscatter. The output factor was measured using a parallel plate chamber ( N233431754, PTW, Germany ) placed in 5 cm of solid water. Percent depth dose A series of percent depth dose (PDD) curves were measured taken f rom various field sizes between 5x5 and 20x20 cm with a water phantom ( 1D Scanner, Sun Nuclear, Melbourne, FL ) with a ADCL calibrated farmer chamber ( Extradin A12 Ion Chamber, Standard Imaging, Middleton, WI ) and associated electrometer ( Victoreens Model 530, Elimpex, Austria) The measurements were taken in step and shoot position intervals of 1 cm from 3 to 10 cm within the phantom, not in the continuous fashion that is common to this type of measurement. Each dosimeter was shielded with a coating to prevent damage from water. For each field size, a PDD graph was created using the raw PDD and the corrected PDD data using the nonparalyzable deadtime model. The data was corrected by iterative methods to find the optimal correction factor that matched the PDD data using the least squares method. The correction factor is denoted as since the count rat e could never be determined for every exposure. This is because the detection system is limited to an integration time of 10 ms. Therefore equation 29 is modified to evaluate for a series of equivalent pulse (magnitude and time profile) that register M total counts and N true counts. 1 = 1 + (2 10)

PAGE 57

57 In Equation 210, the parameter is equal to / where t is the equivalent time that corresponds to the average count rate from M measured counts. Results Gadolinium Oxysulfide Scintillator vs. the Plastic Scintillator Tables 22 and 23 contain the recorded signal and reference counts from the two dosimeter s as well as the calculated output factors and net counts for field sizes of 3x3, 5x5, 10x10, and 15x15 cm. T he net plastic counts decreased with increasing field size which was unexpected given the higher output factor. This led to the measured output fa ctor decreasing with field size when normalized to a standard 10x10 field. The reverse effect was observed in the gadolinium d osimeter which signal fiber and reference fiber responded proportionally to changes in field size. The importance of these results indicate how well the signal counts correlate with the reference counts in the gadolinium scintillator over all field sizes, where w ith the plastic scintillator, the proportion begins to saturate after 10x10 cm significantly. The gadolinium dosimeter measures the output factor very well below 10x10 cm, but over under responds at field sizes greater than 10x10 cm. Figure 2 4 illustr ates the effect of the pulse repetition rate. Although the instantaneous dose rate isnt changing, at higher frequencies counts from one pulse may start to overlap with counts from next pulse. The magnitude of this effect would have been expected from the gadolinium oxysulfide scintillator given i t s longer decay time. However, the plastic scintillator demonstrated higher repetition frequency dependence. Over all, the effect appears to be negligible given the magnitude of the variation in response. The overall trend appears to be linear with fractional losses

PAGE 58

58 respectively. It should also be noted that the sensitivity between the two dosimeters is vast as the gadolinium was 30 times more sensitive to radiating despite a much lower detector mass and equivalent coupling efficiencies. Stem Count Behavior Data was acquired from various radiation sources (sources A [6MeV electron], B and D ) delivering variable quantities of radiation to the stem of the d osimeter by varying the field size. The scintillator, or sensitive element, was shielded to isolate the stem counts. The variable length was achieved by adjusting the field size on each of the sources and providing limited backsc atter material to eliminate field effects. The counts from each source were normalized to the maximum field size reading to create a plot of normalized stem counts vs. field size. This representation is illustrated in Figure 2 5. The slope of the x ray sources and proton source is comparable (0.038 relative counts/cm vs. relative 0.0329 counts/cm). Note that the protons energy is quantified in terms of the range of the particle. 14 cm proton is the nominal energy of clinical proton therapy with values ranging from 5 to 23 cm. The 1000 MU per min PRF was included to illustrate an observed effect when the PRF was set very high the counts would decrease, although this effect didnt happen all of the time as in Figure 24. System Time Resolution Continuous s ources Radiation sources B, C, and D were used to evaluate the PSD dead time from different particle types and energies. The data was fitted by placing the measured inverse count rate data (y axis) along with the inverse dose rate (x axis) and fitting the data to a onedegree polynomial function Figure 26 graphs the data collected from the polystyrene scintillator irradiated from sources B, C, and D. Calculated dead times from

PAGE 59

59 Table 24 indicate that dead time is affected by the radiation source and energy of the ionizing particles. The measured dead time increases from the induction of Cerenkov radiation into the background fiber. It is also noteworthy that the slope of the fitted functions that corresponds to the sensitivity when corrected for dead time of the cobalt 60 source and the 14 cm proton source is 9.78 million counts per Gy and 10.1 million counts per Gy respectively. The nonparalyzable model is an excellent representation of the count rate data and very useful for calibrating the dosi meter for sources that are known to be energy independent for polystyrene. Pulsed x ray s ource Dose rate effects were analyzed using the PSD system by varying the source to detector distance (SDD) to generate a series of dose rates for a 6MV photon beam. The measurements were preformed free in air using an ionization chamber with buildup cap to measure the charge collected at each position along the beams central axis. The PSD measured scintillating counts by subtracting the signal counts from the refer ence counts. Figure 27 A shows the dosimeter counts vs. charge collected by the ion chamber. The more charge collected indicates a higher dose rate since the dosimeter was positioned closer to the source. Charge greater than 6.7 nC resulted in a decrease which is not indicative of nonparalyzable deadtime behavior as shown in Figure 27 B. The dose rate cannot be quantified due to the unknown measurement time of both dosimeter s, therefore the total charge was considered to be proportional to the measured dose rate. The counting times in both cases could be regarded as the same since the exposures were kept consistent at 100 MU at 400 MU/min. The y intercept was calcu lated using the data that was lower than the threshold as shown in Figure 27 B. The intercept ( ) is shown in the last row of Table 25.

PAGE 60

60 Percent Depth Dose Curves Figure 28 compiles the four PDD curves from the 5x5, 10x10, 15x15, and 20x20 cm field sizes for the raw, corrected, and actual PDD data. The corrected PDD data were calculated by subtracting the background counts and using Equation 27 except instead of count rates, restricting m to counts detected and the dead time to a correction factor that accounts for the dead time divided by the counting time to find the count rate. This was done because without proper equipment it is possible to determine the average or instantaneous count rate. The dosimeter was exposed to 100 MUs of x rays delivered at 400 MU/min (same as in the previous experiment) consistent throughout all measurements. The PDD curve was measured along the beams central axis at 100 SSD to the water phantom. The data was normalized to the 3 cm measurement location. The value of used to correct the raw data is shown in Table 25. Random Phenomena When performing measurements at high photons and electron using the PSD system random spontaneous pulses were detected by PMT. These occurrences were random and independent of the detector. Figure 29 illustrates the effect of a 6MV photon beam performing PDD measurements from the previous experiment. On occasion, an isolated high count rate pulse w ould be detected by one PMT but not the other. These events would appear frequently and the measurement would be discounted. Figure 210 is a count rate plot from two different experiments illustrating the sampled counts at 500 ms at increasing optical l ength exposed in the cobalt irradiator. After a certain count rate threshold the counts would decrease despite more optical in the irradiator. In both cases the counts in PMTs would drop. These instances

PAGE 61

61 demonstrate the unusual behavior of the detection system in a high energy radiation environment and may provide further insight into the experimental counting issues associated with the PSD detection system. Discussion The background subtraction is a proven and useful method for correcting the stem re sponse in the radiological and proton therapy detection applications; however is an inadequate method in radiotherapy CLINAC dosimetry using this PSD detection system The effects of dead time along with utilizing these correction methods in radiological imaging dosimetry will be examined in the subsequent chapter. The other studies that used the background subtraction method experimented with PMTs equipped or other photon detectors use charge integration or operate their detectors in current mode. Detectors that are applied to radiation dosimetry at high event rates are normally operated in current mode. The PMT system in this study contains an amplifier, comparator, pulse shaper, and counter system designed to digitize signal and transmit directly to a personal computer The detector operates in pulse mode which has shown in this study to make it susceptible to pulse overlap and loss of information. The nonparalyzable model in general is an acceptable model for making proper corrections to obtain accurate dosimetry for this dosimetry system however is limited in electron/x ray radiotherapy by the instantaneous dose rate. Dead time corrections may play a future key role in proton therapy applications for dosimetry measurements in the Bragg pe ak region of the proton PDD curve. Publications have cited a linear energy transfer (LET) dependence that using this model may be partially correctable.8183 The charged particle LET often described in terms of the linear mass stopping power, can be corrected for ionization quenching by analyzing the energy deposition rate (dE/dt)

PAGE 62

62 which can be quantified readily in counts per second by the detection system. Both quantities increase at lower proton kinetic energies however in varying magnitude since dE/dx is proportional to one over the particles velocity squared and dE/dt is proportional to one over the velocity. Scintillation counting systems especially the one investigated in this study are limited in their ability to capture the entire signal at high count rates. This is because of deadtime effects that prohibit the system from distinguishing between pulses separated by a finite width. It is quite evident that the dead time of the plastic scintillator is increased when exposed to radiation capable of producing Cerenkov radiation. This is thought to be directly tied with the transient absorption effects happening within the optical fiber. The gadolinium dosime ter proved to be excellent at overcoming the field size effects producing accurate results in the smaller field sizes where scatter is less prevalent. The success of the gadolinium dosimeter was att ributed to its long decay time small length, and emission spectrum. The dose rate experienced by these dosimeter s from radiation delivered from a LINAC using the values from Table 21 is anywhere between 25 to 100 Gy/s (1,000 rads/s) which is very high when compared to the other clinical modalities. However the results from the cobalt experiment prove that even at lower dose rates the quenching or extended dead time is present. The data by Mattern show s that delivery from pulsed radiation significantly increases the absorption of light photons within an opt ical fiber owing to a transient recovery of defects. These transient effects contribute to the dead time of the entire counting system by affecting the time that no scintillation light can be transmitted after the optical fiber is irradiated. Due to th e short decay time of polystyrene, the fraction of

PAGE 63

63 counts reaching the PMT was determined by the amount of luminescence created within the optical fiber. However, the pulsing of the radiation source clearly has a strong effect causing it to deviate from the nonparalyzable model at high dose rates and pulse rates. It also demonstrates a maximum count rate threshold that will cause even more losses if the count rate exceeds the maximum. This was consistent with the results presented by Daams. It still r emains unclear why the stem count rate plays such significant role in increasing the dead time. Further analysis such as the scintillator time intensity profile and the emission spectrum of the scintillator under a diagnostic beam compared to radiotherapy may provide further insight. Conclusion This investigation has led to two very important results. First, the background subtraction method is an acceptable method for low energy x ray dosimetry for applications when Cerenkov production is not a concer n. From this result, the correction factor may be found by the methodology discussed in this which is based on correlating a linear response to varying field sizes between the signal and reference. fiber. This method is superior than just solely measur ing a PMT correction factor as it accounts for all the parameters that contribute to the loss of signal. The second and probably most impactful conclusion is the effect of dead time in high energy scintillation counting experiments. However, in most of t he cases it was correctable using the nonparalyzable model correction method. When constructing the ideal PSD it should have a long temporal response and emit light with a similar wavelength as the gadolinium oxysulfide scintillator It should also be characterized for dose rate dependence and not rely on the pulse repetition performance as a means of evaluating count rate effects since they are different.

PAGE 64

64 Ideally, with the proper equipment these measurements could be potentially investigated to analyze the li ght spectrum of the gadolinium and PSD as well as profile the emitted pulse temporal response.

PAGE 65

65 Table 2 1 M easured pulse width at half maxi m um and pulse period for the Varian 69 Average dose rate (MU/min) Fraction of pulses Period (ms) Dose per pulse (Gy/pulse) 600 1 2.77 2.78 x 104 500 0.8 2.77 2.78 x 104 0.2 5.55 2.78 x 104 400 0.5 2.77 2.78 x 104 0.5 5.55 2.78 x 104 300 1 5.55 2.78 x 104 200 1 8.33 2.78 x 104 100 1 16.66 2.78 x 104 Table 2 2 Plastic counts from signal and reference fibers as well as output factors normalized to a 10x10 cm field size from a 50 MU irradiation. Field Size (cm x cm) Signal fiber (#) Reference fiber (#) Net Counts (#) Output Factor (measured) Output Factor (accepted) 3 x 3 48210 25444 22766 0.302 0.681 5 x 5 86481 47575 38906 0.517 0.793 10 x 10 178801 103512 75288 1.000 1.000 15 x 15 180610 127688 52921 0.703 1.343 20 x 20 182008 136519 45489 0.604 1.477 Table 2 3 Gadolinium counts from signal and reference fibers as well as output factors normalized to a 10x10 cm field size from a 50 MU irradiation. Field Size (cm x cm) Signal fiber (#) Reference fiber (#) Net Counts (#) Output Factor (measured) Output Factor (accepted) 3 x 3 1551329 155107 1380711 0.691 0.681 5 x 5 1785994 201427 1564424 0.783 0.793 10 x 10 2288194 263002 1998892 1.000 1 .000 15 x 15 2684579 301682 2352729 1.177 1.343 20 x 20 2997395 331994 2632202 1.317 1.477

PAGE 66

66 Table 2 4 Calculated dead times from various radiation sources Scintillator Radiation source Gadolinium Orthovoltage 82.3 Gadolinium Cobalt 60 698 Polystyrene Orthovoltage 629 Polystyrene Proton 685 Polystyrene Cobalt 60 5356 Table 2 5. Calculated dead time parameters of a 6 MV photon beam free in air and percent depth dose (PDD) measurements. Measurement technique Dead time parameter [ ] (unitless) PDD field size: 5 x 5 3.10E 05 PDD field size: 10 x 10 1.01E 05 PDD field size: 20 x 20 1.13E 05 PDD field size: 20 x 20 1.10E 05 Free in air 1.05E 05

PAGE 67

67 Figure 21. Schematic illustrating the acceptance cone of the optical fiber in relation to the Cerenkov emission cone of the electron transversing through the fiber. Figure 22. Pulsing kinetics of a typical LINAC X rays and electrons are delivered in a series of pulses which are actually a set of smaller pulses of radiation released in sync with the magnetron. CLADDING CLADDING CORE C electron

PAGE 68

68 Figure 23. PMT time response from a finite light pulse at the photocathode. Time Rate of electrons arriving at the anode Electron Transit Time Transit Time Spread (TTS)

PAGE 69

69 Figure 24. Pulse repetition rate effects on the dosimeter s response. PRF values correspond to pulse frequencies ranging from 2.77 ms up to 16 ms. These frequencies are represented graphically as 100 MU/min to 650 MU/min. 100 200 300 400 500 600 700 0.97 0.975 0.98 0.985 0.99 0.995 1 1.005 Pulse repetition rate (MU/min)Normalized response Plastic Gadolinium scintillator

PAGE 70

70 Figure 25. Optical f iber r esponse to varying lengths of radiation exposures delivered in a series of field sizes to the optical fiber. A linear fit was applied to the data to analyze the linearity of the stem signal generated in the background fiber. E rror bars are shown to represent 1 standard deviation from the mean. 0 5 10 15 20 25 30 0 0.2 0.4 0.6 0.8 1 1.2 Radiation field length (cm)Normalized response 14-cm protons 120-keV x-rays 6-MeV electrons at low PRF 6-MeV electrons at high PRF

PAGE 71

71 Figure 26. Plot of inverse count rate and dose rate to measure dead times from cobalt 60, orthovoltage (350 kVp x rays) and clinical protons (12 cm range in water) used in proton r adiotherapy irradiations Data fitted to a linear model to calculate the y intercept. Dead time values are listed in Table 24. Error bars are included and represent 1 standard deviation from the mean. 0 20 40 60 80 100 120 0 0.5 1 1.5 2 2.5 x 10-5 Inverse dose rate (s/mGy)Inverse count rate (s/counts) Cobalt 60 Proton Orthovoltage

PAGE 72

72 Figure 27. Megavoltage beam irradiation free in air. Top figure ( a ) shows the net dosimeter counts from different dose rates. The bottom figure ( b ) is those data arranged to fit to the nonparalyzable dead time calculation. The dose rates were obtained by moving both the dosim eter and the ionization chamber along the gantry axis. 3 4 5 6 7 8 9 10 4 5 6 7 x 104 Charge collected (nC)Response (counts)a) 0.1 0.15 0.2 0.25 0.3 0.35 1 1.5 2 2.5 x 10-5 Inverse charge collected (1/nc)Response (1/counts)b)

PAGE 73

73 Figure 28 Percent depth dose curves for field sizes 5x5 cm ( a ) 10x10 cm ( b ) 15x15 cm ( c) and 20x20 cm ( d ) containing the measured curves using the ion chamber, PSD raw data, and PSD corr ected data. Red markers indicate the raw data, green the corrected data, and blue is the measured data using an ionization chamber. PDD was normalized to the 3cm measurement depth. 2 4 6 8 10 0.8 1 1.2 1.4 Depth (cm)PDDa) 2 4 6 8 10 0.7 0.8 0.9 1 Depth (cm)PDDb) 2 4 6 8 10 0.7 0.8 0.9 1 Depth (cm)PDDc) 2 4 6 8 10 0.7 0.8 0.9 1 Depth (cm)PDDd)

PAGE 74

74 Figure 29. Spontaneous pulse from the detection system wit hin the signal optical fiber but not present in the reference optical fiber. Each data point corresponds to the counts sampled by each PMT at 50 ms. 0 2 4 6 8 10 12 14 16 0 0.2 0.4 0.6 0.8 1 Elapsed time (s)Normalized counts per sampling time signal detector reference detector

PAGE 75

75 Figure 210. Counting profiles from incremental insertion of the PSD into the cobalt irradiator. Both counting profiles (blue and green) are captured from various amounts of optical fiber irradiated. Light color line is the signal optical fiber and the dark color line is the reference fiber. After the maximum count rate (normalized to 1) the counts began decrease despite more detector exposure. 0 20 40 60 80 100 120 0 0.2 0.4 0.6 0.8 1 Elapsed time (s) Normalized counts per sampling time Trial 1, signal fiber Trial 1, refernce fiber Trial 2, sigal fiber Trial 2, refernce fiber

PAGE 76

76 CHAPTER 3 DEVELOPMENT OF AN ENGERY INDEPENDENT FIBER OPTIC DOSIMETER FOR RADIOLOGICAL DOSIMETRY Energy Dependence in Radiological Dosimetry The term energy dependence is used to describe the inaccuracy of using a dosimeter caused by the energy of the incident radiation, to measure absorbed dose from ionizing radiation. In the low energy range of medical x rays, it is often attributed to the difference between the mate rial composition of the dosimeter and the evaluated biological tissue.84 Variatio ns in density, atomic weight, and effective atomic number lead to differences in photonenergy absorption resulting from more or less photon interactions which deposit energy inside the sensitive element. Diagnostic energy x rays deposit a majority of their energy through photoelectric absorption making the response of the dosimeter highly dependent on the effective atomic number (Zeff) of the sensitive material. The mass en probability of photon interacti on that results in energy transfer to localized charged particles. Since the coefficient is inherently energy dependent, the effective mass energy absorption cross section is calculated by weighting each energy component based on the incident photon energ y spectrum. Further complicating the issue, t hese coefficients change drastically with respect to energy in the diagnostic energy range. It extremely difficult to directly measure the ratio of mass energy absorption coefficients ; which is why dosimeter s need to be calibrated under their measurement conditions. The reference dosimeter used for calibrating is an ionization chamber, which is known to be the gold standard in radiological dosimetry. D iagnostic ionization chambers show a variation in energy response of 5% from 1 keV up to 1.33 MeV .85 They produce the most reliable measurements making them ideal candidate to calibrate other types of

PAGE 77

77 dosimeters However, they are used primarily for machine calibration and quality assurance. Table 31 taken from the International Atomic Energy Agency (IAEA)85, lists the recommendation specifications of a ref erence class for various imaging modalities. The disadvantage of ionization chambers is their size which makes them unsuitable for measurements in phantoms that do not already accommodate these chambers. In order to perform patient dosimetry using phant oms, it would be ideal to use a dosimeter that is small and compact, as well energy independent, all while preserving the other characteristics that are desired in radiation dosimetry Currently there exist a select group of solid state dosimeter s that have been characterized for clinical radiological dosimetry .8688 These are: thermoluminescent dosimeter s (TLDs ) optically stimulated luminescent dosimeter s (OSLs), and metal oxidesemiconductor fieldeffect transistor s (MOSFET s). The f irst two are crystal structures able to retain a fraction of the absorbed energy within metastable states that result from defects in their material. A crystallized lattice structure, stores energy in trapped electrons that maintain this metastable state until they are excited enabling them to recombine w ith neighboring electrons and release energy in the form of light or heat. The amount of light emitted and collected is proportional to the amount of energy absorbed. The two methods of exciting the trapped electrons within the crystal lattice structure are in the form of heat, thermoluminescence, and light, optically stimulated luminescence. The dual MOSFET dosimeter is a type of p n junction semiconductor that is two diodes on the same chip operating at two different gate biases. This architecture produces more signal and allows for real time monitoring. M ore stability is gained and minimal temperature/pressure corrections are required.

PAGE 78

78 Thermoluminescent Dosimeter s TLDs are an acceptable dosimeter for performing patient dosimetry. A review of us ing TLDs for dosimetry in diagnostic radiology was done by Zoetelief e t al.89 Thermoluminescent crystals are very sensitive to radiation making them exceptional low level r adiation detectors. Th ey are especially effective when fabricated in the form of loose powders, disks, squares, and rods. Only a small amount of material is needed to form the sensitive region. They require an extensive amount of preparation and handling to avoid fading and annealing processes that affect measurement. For low energy dosimetry the sensitive element material is an inorganic compound, Lithium Fluoride (LiF), with Mg, Cu, P, Ti activators. In particular LiF:Mg:Ti (TLD 100) has been commonl y employed in clinical dosimetry because of its near tissueequivalent effective atomic number, 8.2, compared to 7.4 for soft tissue. LiF:Mg:Cu:P (GR 200) is another type of TLD material that offers 30 to 50 times more sensitivity and exhibits less supral inear behavior than Td100, although has demonstrated a more significant energy response to low energy x rays. Response in sensitivity for LiF:Mg:Ti has been characterized within 9% over diagnostic x rays ranging from beam qualities of 1.48 to 6.33 mm Al .88 TLDs are not ideal for clinical diagnostic x ray beams due to an extensiv e annealing process which desensitizes the crystals and a nonlinear dose response (supralinearity) over diagnostic energies. Optically Stimulated Luminescent Dosimeter s Optically stimulated luminescence dosimeter s (OSLDs) offer excellent sensitivity to radiation as well as are manufactured with a small size making them very useful for in phantom dosimetry. Over the past few years, they have started to replace TLDs as the most popular instrument for radiological d osimetry because of superior

PAGE 79

79 performance.90,91 The sensitive element of a NanoDot OSL dosimeter is Al2O3, effective Z of 11.3, which causes an over response to low energy photons as a result of high contribution from the photoelectric effect. The sensitive volume consists of a cylinder 0.5mm in diameter that extends to 0.1 mm in length. As a result of their small size good spatial resolution is obtained. OSLs are calibrated to measure point doses using a small volume ion chamber. TLD and OSLD light output is measured using a light photon counting device such as a photomultiplier tube (PMT) Commercially available dosimetry devices employ OSL technology to create small compact dosimeter s for a variety of dosimetry applications. In terms of energy dependence, 15 % variation was observ ed by one study over CT x ray beams of 80 and 120 kVP.92 Another study evaluated the energy dependence over the entire diagnostic range and a vari ation of 33% across beam spectrums with effective energies from 25 to 45 keV .93 M etal Oxide Semiconductor FieldE ffect Transistor Like the dosimeter s described above, MOSFET s have a relatively small size, which aids in their ease of use. They produce instantaneous measurements making them a lot faster than the OSLD and TLD.94 The limitation of using MOSFET s is that they do not last as long/are not as reusable as other dosimeter s leading to a long term cost due to progressive loss in s ensitivity MOSFET dosimeter s have demonstrated undesirable reproducibility fluctuations that gradually occur at low keV energies over long time periods. This requires MOSFET s to be recalibrated after a certain amount of usage.95 Their energy correction factor has shown to approximatel y double going from a 30 kV x ray source to a 150 kV source.87 S tudies have concluded that the under response in lower keV energies, MO SFET s may be the least ideal for low energy photon dosimetry.87, 95

PAGE 80

80 Tissue Equivalent Fiber Optic Dosimeter Background Plastic scintillating dosimeter s (PSDs) are a promising dosimetry tool for low energy photon dose measurements They have been strictly utilized in researchbased studies, and have yet to enter into the commercial marke tplace. PSDs are highly sensitive, exhibit good counting statistics and demonstrate approximate tissue equivalence in the diagnostic energy range ( less than 150 keV). PSDs produce instantaneous measurements allowing for real time dosimetry and require no post irradiation processing. Ideally, the response of a PSD relative to soft tissue would remain independent of energy, direction, and dose rat e Studies that have characterized PSDs have observed them to be susceptible to these effects.78, 96 97 Angular dependence is attributed to misalignment between the sensitive element and the optical waveguide. However from a dosimetry perspective, placing the dosimeter in a scattering medium significantly minimizes this effect .98 The most concerning issue of using PSDs for medical x ray imaging dosimetry is relating the response of the dosimeter to the difference in energy absorption properties between biological tissues and plastic. Typically PSDs used for radiological exposures are calibrated to the energy of the x ray spectrum to mimi c the response of air filled gas detectors which have been shown to resemble soft tissue. The energy dependence of the PSD is also manifest as a variation in sensitivity as a function of depth in tissue equivalent material that has been characterized as high as 10% across 16cm of solid water for a 120 kV beam.78 Variation in sensitivity has been attributed to beam quality changes induced by beam hardening and scatter. Lessard et al.99 developed a systematic approach for obtaining correction factors relating the absorbed energy per unit mass of polystyrene to water to

PAGE 81

81 offset the energy response of polystyrene for x ray beams characterized with an effective energy less than 86 keV. For the diagnostic x ray sources examined, it was demonstrated that the correction factor was directly correlated to the spectrally averaged mass energy absorption coefficients of polystyrene and water demonstrating the application of large cavity theory to radiological dosimetry. Studies that have investigated other d osimeters with less air equivalence t han plastic scintillators, have noted that the response at various depths have shown good agreement with ionization chambers.94, 100 The s imilarity in response despite differences in material composition of each dosimeter s sensitive volume may prove dept h depen dent correction factors are nonvarying due to little change in the incident x ray spectrum. This evidence suggests the role of quenching in plastic scintillators, which has been characterized below 125 keV, may play a larger role in depthdose meas urements due to lower energy contributions from scattered photons.101104 The effect of quenching was extensively studied by Frelin who investigated the response of plastic scintillators to low energy photons and electrons, and observing the linearity of plastic scintillators deviated under 200 keV for photons and 100 keV for electrons. Two S cintillators Compacted i n to a S ingle D osimeter P ackage Hyer et al.78 and Moloney98 have characterized polystyrene and gadolinium oxysulfide (GOS) fiber optic dosimeters respectively for radiological dosimetry. In their respective investigations they have observed that both dosimeters demonstrate positive energy dependence with respect to tube potential. A correlation from the ratio of the res ponse between the two dosimeters would prove to be extremely useful for correcting the depth dependence of the PSD, especially due to the sensitivity of GOS scintillator to scatter radiation. In theory, changes in sensitivity of the ratio of plastic count s to

PAGE 82

82 gadolinium counts with depth would directly correlate with the change in sensitivity of the PSD. This manuscript characterizes the effectiveness of using a dual scintillator dosimeter to resolve the energy dependence of the PSD Due to the high sensi tivity of GOS to x rays, the scintillator needs to be evaluated for dead time, in a similar fashion explained in Chapter 1, however using a radiographic x ray tube. Table 3 1 tabulates the physical properties between both scintillators including the effec tive atomic number, decay time, and rise time. Note the differences in decay time between the two scintillators. Materials and Methods Construction of the Dosimeter The dosimeter was constructed as a package c ontaining three cylindrical 500micron diameter multi mode optical fibers (Eska CK 20, Mitsubishi Rayon Co, Tokyo, Japan), two of which were coupled to scintillators, with the remaining fiber left uncoupled to quantify radiation induced fluorescence as described in Chapter 2. This reference fiber was positioned adjacent and parallel to other fibers to quantify all stem counts originating in the scintillator waveguide. These stem counts were subtracted from the measured signal in the optical fibers The polyst yr ene scintillating elemen t ( BCF 12, St Gobain Crystals) was manufactured as a cylindrical fiber of similar dimension (500 microns) to the optical fiber. T he coupling surface of each fiber was polished using the technique outlined in Chapter 1. The scintillator was optically coup led to the optical fiber using a transparent adhesive gel ( Krazy Glue ) and 1 cm of 0.12 cm heat shrink tubing (Fisher Scientific) to create tighter coupling. After coupling, the opposite end of the plastic scintillator was cut to the desired length of 3m m and coated with the reflective coating.

PAGE 83

83 The GOS scintillator was obtained in the form of a 100 m layer from a rectangular sheet of an image intensifying screen for clinical screen film mammography (Kodak MIN R 2 Cassette, Eastman Kodak Company) The GOS was coupled to a polished optical fiber by applying a thin coat of the adhesive gel to the polished end of the optical waveguide and placing the fiber onto the slab while exerting pressure to enhance the binding process. After a short time, the optic al fiber was removed along with the attached thin layer of GOS coupled to the end of the fiber. To provide structural support and shielding from ambient light, each scintillator fiber element and the reference fiber w ere encased in heat shrink tubing to c reate a single package containing all three fibers. A single PMT is required for monitoring each scintillator element Therefore, the fibers were branched off from one another to allow for a n individual connection with the PMT array. The uncoupled end of each optical fiber was fitted with a female SMA 510 connector (SMA 905) using a SMA epoxy (FiberFin, 1656 resin, #80 hardener). The remaining length of uncovered optical fiber was wrapped with heat shri nk tubing. A male SMA optical fiber adapter (E577651, Hamamatsu Corporation, Bridgewater, NJ ) for each PMT was integrated within the protective PMT housing unit to interface each fiber to the PMT. Maintaining a small size of the dosimeter is important for dosimetry measurements involving anthropomorphic phantoms, since the size is an important factor in how easily the dosimeter can be integrated with the phantom. Figure 31 illustrates the size of a completed dosimeter relative to a U.S. quarter. The photon detection system, used by Hyer78, was designed for low level photon counting, using an array of PMTs each mounted with a microcontroller that permits data

PAGE 84

84 transfer. C ounting data was buffered and transferred through a hub to a computer for readout. A USB/serial hub (UPORT 16108, Moxa Inc, Brea, CA) provided a pathway to transmit data between t he computer and the PMT array. Data acquisition settings and communication between the computer and PM Ts were controlled by a custom MATLAB (MathWorks, Natick, MA) graphical user interface program permitting the user to monitor counting data, PMT gate time, as well as graph and save all acquired data. Determining the S ystem Time R esolution and S cintillat ion E fficiency The time resolution, or dead time of the dosimeter was evaluated by performing a series of measurements with a variable exposure rate to the scintillator, which was achieved by varying the x ray tube current. To measure the true count rat e a clinical CT thimble chamber ( Radcal RC0.6CT, Monrovia, CA ) was used to quantif y the exposure rate and exposure time. This ionization chamber was described in AAPM Report No. 111 as an appropriate instrument for CT dosimetry.105 The observed count rate and exposure rate were calculated by averaging the net scintillation counts or exposure and dividing by the exposure ti me. For this method, the true count rate is expressed in terms of the exposure rate, as the two are related by a conversion factor. The dosimeter dead time was calculated by expressing the data in terms of the inverse observed count and the inverse expo sure rate stated in Equation 29 T he dead time can be solved for using linear regression to find the optimal y intercept of the inverse of the true count rate and observed count rate data: 1 = 1 + (3 1) In E quation 3 1 the theoretical value of n is the product of the exposure rate and a conversion factor relating the number of true detected counts to the measured

PAGE 85

85 exposure rate. The two are related by the absorbed dose rate to the scintillator: = / / = (3 2) Equation 32 is valid under chargedparticle equilibrium and the condition that the Roentgen is measurable; otherwise, the expression is invalid since the relationship between dose and exposure breaks down. The 0.51mm heat shrink tubing that encases the scintillator provides sufficient buildup to achieve chargedparticle equilibrium for diagnostic beams as the CSDA range of a 150 keV electron in polysty rene is 0.3 mm.106 is defined as the energy absorbed by the scintillator per detected light photon and is further examined by Beddar et al.107 The term is the mass of scintillator material in the dosimeter Combining terms the theoretical definition of the conve rsion factor relating the true count rate and the dose rate may be written: = / / (3 3) Equation 33 provides the conversion factor betw een the true count rate and measured dose rate in air To calculate the absolute value of n is unnecessary, since the y intercept solution is dependent only on the relative value of n Therefore when calculating n may be expressed in terms of exposure rate and as long as the energy spectrum remains unchanged, the conversion between the true counts and the ion chamber reading remains constant. The term may be derived from fitting the data to Equation 31. With the use of Monte Carlo techniques, the spectrally averaged mass energy absorption coefficients can be calculated to provide a value for the energy to light conversion efficiency per unit mass for the investigated energies. From this

PAGE 86

86 information the variation in Ssys/m may be evaluated for each investigated tube potential. Exposure Measurements Free in a ir m easurements A CPI Indico Plus RF high frequency xray tube (Communications & Power Industries International Inc., Georgetown, Ontario, Canada) with a variable tube current output was used to deliver a series of different radiological x ray spectrums The dosimeter was positioned 50.8 cm from the tungsten target to maximize the incident exposure rate. The x ray tube current was varied to produce a range of exposure rates irradiating the dosimeter (25 500 mA), while maintaining the tube current timeproduct (25 mAs), field size (17 cm x 17 cm) and beam energy. The exposure rate was measured using the ionization chamber placed adjacent to the dosimeter At each exposure, the net scintillating counts, exposure, and exposure time were recorded. Tube potenti als of 140, 120, 100, 80, and 60 k eV were used to investigate the effect of photon energy on the time response characteristics of the detection system. CTDI phantom m easurements The dosimeter was exposed under varying layers of Polymethyl methacrylate (P MMA) by using multiple hole locations (1.66, 9.67, 16.0, 22.35, and 30.35 cm) on a PMMA CTDI head phantom nested within a body phantom (Radcal, TO CTDI, Monrovia, CA) Using a clinical Toshiba Aqiulion ONE volumetric CT scanner (Toshiba America Medical S ystems Inc., Tustin, CA) radiation was delivered with the tube parked at the 12 oclock position. D epthdose measurements were acquired at all four tube voltages: 80, 100, 120, and 135 kVp using the large focal spot and medium bowtie filter at 200 mAs tube current time product for 1 second with a collimation thickness of 16cm

PAGE 87

87 The surface of the phantom was positioned 40 cm from the focal spot. The ionization chamber was used to record reference air kerma measurements at each hole location. A t each measurement location, the center of the ionization chamber and dosimeter were centered with the central axis of the fan beam. Five measurements were recorded at each location and all tube potentials to generate a calibration curve that would be use d to correct the plastic scintillator reading as a function of depth for each tube voltage. Angular d ependence Due to the architecture of the dosimeter there is lack of symmetry within the sensitive element from the orientation of the plastic and gadolinium scintillators. As a consequence, the angular direction of the radiation may affect the response ratio from perturbation of the x ray beam Therefor e the angular dependence of the response ratio of the gadolinium signal to the plastic was evaluated using the CT scanner which provides fixed beam angle expos ures Static beam angles delivered exposures at 120 kVp and 100 mAs at the dosimeter which posit ioned at the machine isocenter freein air with the table removed from the beam. Two sets one measurements were where the gantry moved one full rotation around the axis of the dosimeter and irradiated in 45 intervals, and another sweep where the dosimeter was rotated 90 and the gantry irradiated the dosimeter normal to axial. In this case 0 corresponded to x ray hitting the tip of the dosimeter unperturbed, and at 180 incident from the back end of the dosimeter and attenuated by the optical fiber. M e asurements were again acquired in 45 intervals.

PAGE 88

88 Calculation of error Performing linear regression using the least squares method to determine the system dead time and fiber sensitivity will induce error and produce an uncertainty associated with each calculation. This will be especially important when testing the dead time of the polystyrene scintillator, which is less sensitive to radiation, making the calculation more uncertain. When performing the regression analysis, three measurements at each of t he six tube currents were taken into account, meaning eighteen data points were included in each dead time calculation. Equations 34 and 35 were used to find the error associated with each parameter of the straight line curve fit characterized by the sl ope and y intercept. ( ) = 1 1 (3 4) ( ) = 1 (3 5) Equation 35 was used to find the error in calculating the dead time. Standard i ) was found by using both error propagation and by modeling the error of each background and signal measurement in accordance with a Poisson distribution. Measurement uncertainties from the ion chamber owing to uncertainty from the standard calibration f actor were (0.7%), effect of beam quality difference between calibration and measurement (2.0%), stem effect (1.0%), and dose rate dependence (2.0%). Considering all of these errors and error from the timer, the uncertainty in any

PAGE 89

89 given exposure rate meas urement was conservatively estimated using the TG 61 protocol as 3.5%.106 The error in generating relative fluence energy spectrums from the Monte Carlo were evaluated by using the simulated statistical error calculated for each energy bin and propagating the error to generate an alternative relative energy spectrum. This spectrum is an equivalent spectrum re presenting the extremes possibilities of the detected energy spectrum. Using this alternative relative energy spectrum the uncertainty in the spectral weighted mass energy absorption coefficients and quenching factors were recalculated to obtain a width of the uncertainty from the Monte Carlo simulations. Monte Carlo Methods Generating p hoton and e lectron e nergy f luence d ata Each scintillator s energy response from x ray absorption properties was evaluated by simulating the initial photon energy spectrum s using Monte Carlo techniques along with stopping power and energy absorption coefficients from the NIST database.108, 109 A source model of each x ra y tube energy output was created as a custom source subroutine within a general purpose Monte Carlo radiation transport code, MCNPX version 2.7 (Los Alamos National Laboratory, Los Alamos, New Mexico). The required probabilistic information regarding beam energy, the shape of the CT x ray fan beam due to the bow tie filter, and x ray field collimation were all experimentally derived from measurements on each machine. The source subroutine for the Toshiba Aquilion ONE was created using the spectrum generat or software with additional information to account the shape of the x ray fan beam due to the bowtie filter and z axis collimation. Methodologies presented in research conducted by Turner et al .110

PAGE 90

90 and Monica G hita111 at the University of Florida were used in concert with a custom MATLAB script function and spectrum generator program to create the source subroutine in FORTRAN 90 ( International Business Machine Corporation, Armonk, New York ). A tool for computationally modeling tungsten anode x ray generators was developed by Siewerdsen based on the TASMIP model of Boone and Seibert.112 113 The software, SPEKTR, allows user to generate tungsten anode spectrum s for different filtrations across a broad range of kVp, filter material, and filter thickness. The equivalent x ray spectrum of the CPI Indico Plus RF high frequency x ray tube was created from three input measurement parameters: peak tube potential and first and second half value layer With the software an initial unfiltered tungsten spectrum is generated. This initial spectrum is iteratively transmitted through different thicknesses of various inherent filtration materials (lead, aluminum, copper, tungsten, and carbon) in 1 keV energy bins using attenuation coefficients from the NIST database for alumin um, graphite, lead, titanium, and copper. T he simulated first half value layer and second half value layer were calculated at each iteration until th e simulated first half value layer is equivalent to the measured for each material. The simulated equivalent spectrum was chosen from the simulated second half value layer spectrums of different filtrations based on the second half value layer that most closely matched t he measured second half value layer With the creation of each source subroutine, the x ray and electron spectra transported through each scintillator can be simulated by modeling the physical experiment. The convenienc e of generating a normalized energy spectrum is that the

PAGE 91

91 quantity of photons generated is not necessary. Spectrum information was captured by binning F4 tallies in energy bins from 1 to 143 keV for photons. Electron energy tallies were compiled f rom 1 to 140 keV in 1 keV increments above 10 keV, 0.5 keV increments between 5 to 10 keV and 0.25 keV bins down to 1 keV Higher resolution energy bins were required in the lower energy levels to account for the large changes in stopping power with energy. For example the difference in stopping power between a 2.25 keV and 2.5 keV electron is equivalent to a 40 keV and 140 keV electron. The custom CT source subroutine was further validated with inscanner CTDI measurements by Long at the University of Florida. First, a free in air measurement was made in the scanner to normalize photon output per mAs, Next, using a R adCal 10cm ion chamber, dose measurements were made at the center and peripheral hole locations of a CTDI head and body phantom. These dose meas urements were than computationally simulated in MCNPX for output comparison yielding an average error less than 3%. The x ray spectrum of both scintillator s was validated by matching both first and second half value layer measurements. Evaluation of i oniz ation q uenching from Birks m odel Once the electron spectrum is obtained, the magnitude of quenching experienced by the dosimeter may be determined. With the assumption that the electron fluence throughout the scintillator will remain unperturbed under el ectron equilibrium and the range of each electron is less than the diameter of the scintillator, each electron of energy E can be modeled to deposit all of its energy into the scintillator Therefore, the light output for a scintillator with scintillation efficiency S from a photoelectron with initial kinetic energy E within the sensitive volume may be expressed using the empirical uni molecular quenching model proposed by Birks and Brooks114:

PAGE 92

92 ( ) = = 1 1 + (3 6) The energy dependent quantity L ( E ), was evaluated at each energy using numerical integration over the entire kinetic energy of the electron. Stopping power information was obtain ed from the NIST ESTAR database, which tabulates electron and positron stopping power data. Published values for the quenching parameter kB for plastic scintillators range between 0.009 to 0.016 cm2MeV1g.104, 115117 A search of the literature found no reported quenching factors for GOS Most inorganic scintil lators that have been evaluated, parameterized the product of kB from Equation 36 between 0.008 and 0.001.76 117 The total quenching factor Q was calculated for each energy spectrum by integrating L ( E ) over the simulated electron energy fluence spectrum and dividing by the integral of the electron energy fluence and the scintillation efficiency product expressed in Equation 37. = ( ) 1 1 + ( ) (3 7) The term S is energy independent and can be factored out of both integrals. All integrals can be evaluated discretely using the initial electron energy fluence data, quenching factor derived from Birks formula, and stopping power data. In Equation 37, E is the secondary electron spectrum, and E is the kinetic energy of the electron as it losses energy within the scintillator Electr on energy spectra for each scintillator were collected in energy bins from Monte Carlo surface tallies. Electron energy tallies were compiled f rom 1 to 140 keV in

PAGE 93

93 1 keV increments above 10 keV, 0.5 keV increments between 5 to 10 keV and 0.25 keV bins down to 1 keV The range of the photoelectrons was found to be less than the diameter of the scintillation volume. The CSDA range for a 125 keV electron is less than 40% the diameter of the plastic scintillator and 1/10th the diameter of the GOS Results Cou nt Rate Performance Light photons originating from the optical fiber and detected by the PMT were measured by placing a reference optical fiber next to the optical fiber carrying the signal from the scintillator The counts from the reference fiber were correlated to counts in the signal fiber by placing the pair of optical fibers under the x ray field and placing the scintillator outside the x ray field and then irradiating different lengths of the stem at multiple beam qualities with tube potentials ranging between 60 to 140 keV. The data was used to generate a correction factor that relates the stem counts detected in the reference optical fiber to the stem counts detected in the signal fiber. Each scintillator demonstrated a l inear re sponse between detected counts in the background fiber and the scintillator fiber. This method was thought to be superior to taking measurements with the reference fiber occupying each PMT because the correction factor could be dependent on param eters other than the PMT efficiency as discussed by Beddar et al .107 The measured count rate data (y axis) along with the air kerma rate (x axis) are plotted for the GOS scintillator and plastic scintillator in Figure 3 2 for the 120kVp x ray spectru m. These graphs show the response of both scintillating elements to varying dose rate. Figure 32 also illustrate s the first degree polynomial fit to the data when presented in the form of Equation 31 to calculate the dead time ( ) and the air kerma

PAGE 94

94 se best fit to Equation 31 are tabulated in the first and second columns of Table II and Table III. Monte Carlo The computed values for the spectrally averaged mass energy absorption coefficients between the scintillation and ionization chamber are tabulated in Tables 3 2 and 3 3 in the fourth column. These ratios were calculated using the normalized x ray spectrum and the mass energy absorption coefficients from the NIST database. The scintillation efficiency for each tube potential was calculated to examine the accuracy of energy correction factors based on the large cavity theory proposed by Lessard and point out any quenching effects from freein air measurements. In the columns to the right, the quenching factors calculated from Equation 37 using the highest and lowest values of the quenching parameter, kB, from the literature. The uncertainty in the gadolinium quenching factor calculation was negligible, while the plastic scintillator was less than half of a percent. The reported statistics by the Monte Carlo run were no more than 0.2% for gadolinium and 1.6% for polystyrene. It should be noted that the absolute scintillation efficiency was not able to be det ermined due to the difficulty in measuring parameters such as coupling efficiencies and transmission losses from the fiber cable. The electron energy spectrum for 120 kV at various positions in the phantom is plotted in Figure 33 The average energy of the spectrum decreases at deep depths due to attenuation of the primary beam and energy contributed from low energy scattered photons. When calculating quenching factors, GOS showed no change in quenching with depth, while polystyrene showed a slight dev iation between 8 and 16 cm but a significant decrease after 16cm.

PAGE 95

95 The photon energy absorption characteristics of both GOS and polystyrene relative to air maintained less than 5% variance between all locations for a 120kVp x ray beam, with polystyrene showing a lesser energy dependence. These results also show that the ratio of the mass energy absorption coefficients between GOS and plastic shouldnt vary by more than 6% across the entire phantom. Due to the small size of the gadolinium scintillator t he x ray energy spectrum at the deepest hole location was statistically insignificant because of a low number of detected photons in each bin. The effect on the plastic scintillator from beam hardening is in fact minimized by the characteristic x ray cont ribution of the GOS which negates the higher energy characteristic x rays of the tungsten anode. These characteristic x rays dominate the energy spectrum of the GOS scintillator leading to a convenient slowly varying response as a function of depth. Fig ure 3 4 illustrates the relative x ray energy fluence spectrums through each scintillator at the phantom top and center hole locations. Energy Response Figure 3 5 plots the normalized sensitivity for each scintillator in terms of counts per Gy as a function of depth into the CTDI phantom for three different tube potentials: 100, 120, and 135 k Vp The data were normalized to the 120 k V p data point at the center hol e location. Each plot includes three sets of data: the GOS corrected for dead time, GOS uncorrected for dead time, and the plastic scintillator uncorrected for dead time. A dead time of 48 ns was used to correct for count losses. This 48 ns value was calculated by using the data from all energies and applying the 1st order polynomial best fit to find the dead time. The data were also examined to find the optimal dead time that would minimize the variance in sensitivity, which was found to be between 50 to 52 ns.

PAGE 96

96 The decrease in sensitivity at greater depths was attributed quenching from the presence of low energy electrons as seen in Figure 3 4. The energy dependence of a dosimeter is often attributed to the variance in sensitivity of the dosimeter fr om the calibration condition. Plastic scintillator response factors are expressed in the quantity net counts per measured air kerma. To determine a correction factor methodology based on a given calibration condition an empirical relationship was evaluat ed to include all CT energies, between the ratio of GOS net counts to plastic counts and ratio of air kerma to plastic counts. The calibration curve is shown in Figure 3 6. The reference condition was the 120 kV beam at the center hole location. The correction factor is equal to the first order coefficient of the calculated best fit using a first degree polynomial with no y intercept function. A noticeable break down in the linearity of the curve was observed for the 80 kVp beam resulting from the ener gy absorption effects as demonstrated by the freein air measurements in Table 3 3 The results show that using this method, the correct plastic scintillator response factor may be estimated to within 9% for CT tube potentials of 100 through 135 kVp and a ll hole locations. Individual percent differences are tabulated in Table 3 4 Each percent difference was calculated by averaging over the 100, 120, and 135 kVp measurements. Column 2 is the percent differences when using the correction method from Figu re 3 4. Column 3 is implementing the same except removing the data from hole locations 1 and 3 and averaging those values separately to get one correction factor. The reason for performing this is because these values fall within the 6% variation in mass energy absorption coefficients and anything outside of this range is due to quenching. The ratio of plastic counts to GOS counts that fell between 0.034

PAGE 97

97 and 0.037 were separated from the regression evaluation since it was deemed their response was independent of quenching effects. Column 4 represents the percent differences by comparing the dosimeter s response to the average of the top and center hole locations without any type of correction method. Angular Dependence Figure 3 7 plots the normalized res ponse ratio of dosimeter for axial and normal axial irradiations. The data was normalized to the 0 measurement on the axial rotat ion exposure. It should be noted that evaluation revealed the effect on the correction factor which is dependent on the rati o of the scintillators response. It is evident that the irradiation from the backend induces a high magnitude of signal loss as well as a more variation in response from the normal to axial irradiation, tha n the axial. This dependence is consistent wit h the PSD and as noted is less of an issue when the dosimeter is placed in a scatter medium which creates isotropic irradiation that is less directionally dependent. Discussion Each scintillator demonstrated count rate nonlinearity effects when compared to the ionization chamber. The behavior of both scintillators can be modeled using the non paralyzable model expressed in Equation 2 7 The results of this model show that on average, the GOS yielded a deviation from linearity of 5% at 26.3 mGy/sec whil e the plastic scintillator demonstrated a 5% deviation from linearity at 324.3 mGy/sec. Exposure rates of clinical CT units are typically lower than 60 mGy/sec demonstrating the impact of correcting for count rate losses of the gadolinium element The re ported exposure rate dependence of the RC 0.6 small volume ion chamber was 2% between 0.0005 to 2033 m Gy/sec well outside the m easured exposure rates that ranged from 2.9

PAGE 98

98 to 333 mGy /sec. However this error in dose rate dependence was accounted for by assigning a fixed error for each ion chamber measurement of 5 %. W hen performing this experiment, a source whose output strength is n o t energy dependent is more ideal for comparing calculated dead times between different energies. The variation in dead time measurements for the GOS scintillator is minimal, and when including all energies the dead time was calculated to be 48 ns. Assigning a dead time value to the plastic scintillator is too arbitrary given the vast uncertainties in the measurements. It is e stimated that the actual value would fall between 80 to 140 ns However presented with an x ray tube with a more intense beam current a precise value could be determined. One reason for not using a megavoltage therapy accelerator or orthovoltage machine was the introduction of Cerenkov radiation above 150 keV. An interesting result is that despite a very long decay time, on the order of microseconds, the GOS element demonstrated a dead time on the order of nanoseconds. According to Knoll, when a pulse of radiation is significantly larger than the dead time of the PMT, which is certainly acceptable to model hundreds of microseconds decay function as a long pulse on the nanosecond second scale, the steady state models may be applied with reasonable accuracy .76 Also noteworthy, as the dose rate was increased, the gadolinium counts were still accumulating after the radiation was stopped insinuating that the decay time increased at higher dose rates. The energy response of the plastic scintillator was most affected at depths beyond 16 cm showing a decrease in sensitivity. This was attributed to ionization quenching effects as the el ectron spectrum is significantly lower as more of the primary beam is attenuated and scatter becomes predominant. The GOS scintillator remained

PAGE 99

99 accurate relative to the ionization chamber among all depths within 13%. The effect of quenching however was l ess apparent in the freein air measurements citing a small variation in the calculated quenching factor. It appears that GOS when properly corrected for dead time, is a feasible dosimeter in the higher radiological energy range along with other high Z s cintillators, and the use of solely a plastic scintillator may require more correction factors. A dual scintillator dosimeter serves an acceptable tool for CT dosimetry demonstrating a superior response to only using a plastic or GOS sensitive element Using the response ratio shows no significant variation from beam direction and high reproducibility. Each scintillating element individually is flawed to some degree, as documented in this manuscript, limiting the minimum energy dependence that can be ac hieved. However, when using both scintillators these flaws can be overcome and the energy dependence can be minimized. Nine percent energy dependence between 100 and 135 kVp is certainly promising and correlating the ratio of the response may even more ac curately correct air kerma measurements into dose to relevant bodily tissues. The biggest advantage of using a dual scintillator dosimeter is that the quenching effect from polystyrene may be minimized and can be potentially used to correct the response o f other PSDs used in an array for rapid dosimetry. With the capability of acquiring data in real time, an analog control program can be coded to measure depthin phantom correction factors for different angular projections of the CT x ray tube as it rotates around the table. Since the dosimeter is fixed, different gantry angles will result in varying depths that the x ray beam must penetrate before reaching the dosimeter The full rotation time of a CT x ray tube is

PAGE 100

100 somewhere between 0.3 to 0.5 seconds. With the detection system utilized in this investigation, counting data can be obtained within 10 ms intervals. If the acquisition time interval on the PMT was set to 50 ms, 6 to 10 depth dose correction factors could be generated in a single axial tube rotation. For helical scans, the number of angular projections correlates directly with the pitch and the PMT gate time. This type of data acquisition technique would be very useful for tubecurrent modulation dosimetry studies. Conclusion This study addressed the count rate effects of using highZ fiber optic dosimeter s for dosimetry purposes, which previously were interpreted as having a strong energy dependence that was attributed to x ray absorption properties. By using a highZ scintillator alo ng with a plastic scintillator, ionization quenching effects may be limited. The dosimeter investigated in this study provides a correction method to offset the depth and energy response of a plastic scintillator used in medical x ray imaging dosimetry st udies. Incorporating this type of photon detection system with fiber optic s allows real time corrections to be made as the gantry rotates around the phantom exposing the dosimeter to the different irradiation conditions characterized in this study. In addition, it proves to be very useful in projection imaging dosimetry where the beam may be highly attenuated when it reaches the dosimeter Although it doesnt perform as well as an ionization chamber, the dosimeter s small size makes it an ideal candidate for physical phantom dosimetry and situations requiring a high spatial measurement without a loss in sensitivity. This would be useful for beam profile measurements or low level background measurements.

PAGE 101

101 After a review of commercially available scintil lators it is recommended that calcium fluoride is the ideal high Z scintillator due to its smaller effective atomic number ( (kB=0.0052).117 However, it is a difficult and expensive process to integrate with an optical fiber while preserving the small diameter. It would be most advantageous to be fabricated into a fiber with a surrounding cladding material similar to the plastic fibers. Scintillators with low energy k edges are ideal because t he k edge will induce a discontinuous relationship when taking the ratio of both scintillators which would make developing a correlation very difficult. The effects of quenching make the dual scintillator detector less ideal for evaluating the beam energy. Although this relationship is relatively linear it still is less than optimal. In the conclusions section of this dissertation, two candidates are brought to attention as possible scintillators that would be most ideal for a dual scintillation detector. The theory of using two scintillators to model the response of biological tissues is very much proven given the proper selection of both scintillators. In addition, the use of plastic scintillators as being more tissueequivalent may be overstated given the gadolinium oxysulfide scintillator out performing the plastic in terms of energy and depth dependence.

PAGE 102

102 Table 3 1. Common properties of GOS and polystryene Property Gadolinium oxysulfide Plastic scintillator (BCF 12) Molecular Formula: Gd 2 O 2 S C 8 H 8 Peak Emission Wavelength (color): 545 nm (green) 435 nm (blue) Decay Time: 3.2 ns Rise Time: Few ns Effective Atomic #: 59.5 5.74 K edge: 50.22 keV < 1 keV Mean Excitation Energy: 493.3 eV 68.7 eV Absorption Photon Yield: Specific Gravity 7.32 g/cc 1.05 g/cc Table 3 2. Polystyrene dead time and correction factor evaluation from free in air measurements and Monte Carlo methods X ray spectrum Scintillator dead time Air kerma sensitivity Mass energy absorption Scintillator dose sensitivity Quenching factor Effective energy (keV) (nGy per count) Ssys/m (106counts per Gy) kB = 0.016 kB = 0.009 39.6 849.6 469.3 821.0 3.10 0.467 0.00014 2.609 0.067 0.720 0.812 47.5 39.46 177.9 695.9 3.37 0.534 0.00010 2.689 0.069 0.737 0.824 53.9 246.7 118.2 637.8 3.87 0.596 0.00012 2.630 0.073 0.746 0.830 59.1 144.2 75.86 618.4 3.90 0.647 0.00011 2.503 0.069 0.750 0.833 62.3 94.36 54.02 601.3 3.98 0.671 0.00015 2.487 0.071 0.756 0.837

PAGE 103

103 Table 3 3. GOS dead time and correction factor evaluation from free in air measurements and Monte Carlo methods X ray Spectrum Scintillator Dead Time Air kerma sensitivity Mass energy absorption Scintillator dose sensitivity Quenching factor Effective energy (keV) count) Ssys/m (106counts per Gy) kB = 0.001 kB = 0.008 39.6 60.75 2.414 2.203 0.044 64.43 0.036 7.045 0.125 0.993 0.947 47.5 54.10 1.244 1.849 0.040 77.44 0.012 6.983 0.149 0.993 0.949 53.9 50.62 0.587 1.726 0.034 83.13 0.194 6.97 0.140 0.993 0.951 59.1 49.65 0.354 0.5890 0.033 82.97 0.324 7.099 0.134 0.994 0.953 62.3 48.58 0.364 1.786 0.0031 79.25 0.387 7.066 0.126 0.994 0.954 Table 3 4. Res p onse factor comparison between estimated air kerma conversion coefficient vs measured Each position averages the percent deviation from the measured for 100, 120, and 135 kV energies. Depth into phantom (cm) Corrected by using all data points Corrected by using selective data points No Correction (response factor determined by middle phantom location) 1.66 5.29% 1.31% 0.38% 1.27% 0.38% 1.27% 9.67 4.90% 2.10% 1.36% 2.03% 5.83% 1.29% 16.0 2.23% 0.72% 0.36% 0.65% 0.36% 0.65% 22.35 2.62% 0.62% 0.84% 0.62% 22.2% 3.30% 30.35 7.85% 1.31% 11.4% 1.33% 247.9% 104.4%

PAGE 104

104 Figure 31. Size of t he dual scintillator dosimeter relative to a United States quarter Photo courtesy of Matthew R. Hoerner

PAGE 105

105 Figure 32. Count rate effects from 120keV x ray exposure to the GOS scintillator (a) and the plastic scintillator (b). The line represents a fit of the data to a second degree polynomial to demonstrate the nonlinearity of the GOS The calculation of dead time for both the GOS scintillator (c) and the plastic scintillator (d) are illustrated using the first degree polynomial best fit. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 x 107 Absorbed dose rate (Gy/s)Gadolinium oxysulfide count rate (#/s)a) 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 x 105 Absorbed dose rate (Gy/s)Polystyrene count rate (#/s)b) 0 10 20 30 40 50 60 70 80 0 0.5 1 1.5 2 x 10-7 Inverse absorbed dose rate (s/Gy)Inverse gadolinium oxysulfide count rate (s/#)c) 0 10 20 30 40 50 60 70 80 0 1 2 3 x 10-5 Inverse absorbed dose rate (s/Gy)Inverse polystyrene count rate (s/#)d)

PAGE 106

106 Figure 33. Electron fluence spectrum through the polystyrene scintillator for various hole locations in the CTDI phantom from exposure to a 120kVp CT x ray beam. Data acquired using Monte Carlo methods. CTDI nested head and body phantom are shown depicting the pos ition of each hole location relative to the x ray tube fixed at the 12 oclock position in the diagram. 0 10 20 30 40 50 60 70 80 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 Energy (keV)Relative fluence Top location Top-middle location Middle location Bottom-middel location Bottom location

PAGE 107

107 Figure 34. Gadolinium oxysulfide (green line) and polystyrene (blue line) relative photon fluence spectrum from a 120kVp CT x ray beam. Hole locations 1 and 3 correspond to 1 (solid lines) and 16 (dashed line) cm respectively. Data acquired using Monte Carlo methods. Note the polystyrene fluence spectrum y axis on the left and gadolinium oxysulfide spectrum y axis on the right. 0 20 40 60 80 100 120 140 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 Energy (keV)Polystryene relative fluence 0 20 40 60 80 100 120 140 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0 20 40 60 80 100 120 140 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 Gadolinium oxysulfide relative fluence Poly (middle hole) Gad (middle hole) Poly (top hole) Gad (top hole)

PAGE 108

108 Fi gure 35. Gadolinium oxysulfide and plastic scintillator energy response as a function of depth into CTDI phantom for three tube potentials: 100, 120, and 135 kV. Three sets of data are plotted to include gadolinium oxysulfide corrected using a dead time of 48 ns, uncorrected gadolinium oxysulfide, and the plastic scintillator. The dashed lines correspond to the +/ 10% deviation from the normalization point. The data were all normalized to the 120 kV center hole data points, which is positioned at 16 cm depth. 0 5 10 15 20 25 30 35 0 0.5 1 1.5 100 keVRelative detector sensitivity in normalized counts/mGy GadOxySul corrected 0 5 10 15 20 25 30 35 0 0.5 1 1.5 2 120 keV Plastic Scintillator 0 5 10 15 20 25 30 35 0 0.5 1 1.5 2 135 keVDepth into phantom (cm) GadOxySul uncorrected

PAGE 109

109 Figure 3 6 Determination of correction factor for the plastic scintillator for energies 80 to 135 kV. Correction factor was calculated by taking the response factor of each scintillator and normalizing to the calibration condition which was defined at 120 kV beam at the center hole location. Solid line represents the linear no threshold fit and the two dashed lines represent the standard deviation of the fit. Top Figure (a) shows the entire range of data. Bottom Figure (b) is the range of data (within the grey box) that were deemed unaffected by quenching. 0 200 400 600 800 1000 1200 1400 1600 0 1 2 3 4 5 6 Plastic correction factor (Gy/counts) Ratio of net gadolinium counts to net plastic counts a) linear no threshold fit standard deviation of fit 200 250 300 350 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 Plastic correction factor (Gy/counts) Ratio of net gadolinium counts to net plastic counts b) linear no threshold fit standard deviation of fit CF = 0.00355*R

PAGE 110

110 Figure 3 7 Angular dependence of the response ratio of the dosimeter to an axial irradiation (a) and normal axial irradiation ( b). In both cases, the data has been normalized to the zerodegree axial irradiation response. The schematic to the right of each plot illustrates the experimental setup with the shaded circle representing a head on view and the square a side view of the scintillator. The white rectangle indicates the optical fiber, which is not visible in the top diagram 0 90 180 270 360 0.9 0.95 1 1.05 1.1 Axial rotation (degrees)Gadolinium-to-plastic response ratioa) 0 90 180 270 360 0 0.2 0.4 0.6 0.8 1 1.2 b) Normal-to-axial rotation (degrees)Gadolinium-to-plastic response ratio 0 270 180 90 0 270 180 90

PAGE 111

1 11 CHAPTER 4 CHARACTERIZATION OF A LINEAR FIBEROPTIC DOSIMETER FOR COMPUTED TOMOGRAPHY Introduction The convention for computed tomography (CT) physical phantom dosimetry is performed with dosimeter s that have a small sensitive volume to satisfy cavity theory specifications and be integrated easily within phantoms In ra diological dosimetry, meeting cavity theory requirements is easily achievable due to a small wall thickness required to provide sufficient buildup to create full charge particle equilibrium and localized charged particle energy deposition. The principle issue in radiological dosimetry is matching t he material composition of the dosimeter to replicate the photon energy absorption characteristics of the biological tissue of interest. Ionization chambers are the gold standard in radiological dosimetry, but are not feasible for internal organ dosimetry due to their large size which limits their spatial resolution and makes them difficult to integrate within phantoms. This has prompted the development of dosimeter s that are small enough to be easily integrated within all types of phantoms. Dosimetry using physical phantoms has been performed using film, or more recently, solid state dosimeter s such as TLDs and OSLs.118 Film is very useful for measuring dose to an area, but is energy dependent and nonlinear. OSLs and TLDs require a post irradiation process to measure the energy stored within each crystal Com mercially available OSLs and TLD s are made with a small sensitive volume suitable for point dosimetry For example, the size of an OSL n anoDotTM ( LANDAUER, Glenwood, IL) is 10x10x20 mm

PAGE 112

112 When investigating dosimeter s that cannot perform absolute measurements, it may be more advantageous to research a dosimeter with a sensitive volume that represents the geometry of the region of interest to obtain spatially integrated measurements. For example, a thin layer of film can be used to simulate the absorbed dose of a layer of epithelial cel ls. In CT dosimetry, dosimeter s are placed within phantoms that have segmented internal organs. Organs are quite voluminous relative to the size of the sensitive volume of point dosimeter s being utilized in radiological dosimetry studies. Therefore mult iple readings are made and averaged to obtain an organ dose measurement The method of delivering radiation in CT which scans in the table translation direction, allows for partial irradiation of critical anatomy. In addition, with the use of tubecurren t modulation, the intensity of the x ray beam is highly variable along direction table translation. These considerations demonstrate the value of a dosimeter capable of measuring cumulative z axis dose in CT dosimetry studies.119 Suc h a detector does exist though, the 10cm pencil ionization chamber. The detector evaluates the axial dose within a cylindrical CTDI phantom, exactly as the methodology discussed in Chapter 1. However, the detector is unsuitable for inphantom dosimetry d ue to its large size, as wells as no formulism for evaluating dose since cavity theory would not be applicable to this detector.84, 120 Fiber Optic Dos imeter s Fiber optics may permit the development of such a dosimeter P lastic scintillators that have been investigated for photon dosimetry are fabricated into miniature cylindrical fibers with a surrounding cladding material. This allows light to travel within the fiber through internal reflection, and function jointly as a detector an d optical waveguide. When the scintillating fiber is placed orthogonal to the direction of

PAGE 113

113 incident radiation, the small diameter of the fiber maintains charged particle eq uilibrium. The ideal dosimeter would demonstrate uniform spatial sensitivity across its active length. There are four main issues when considering the spatial sensitivity of a scintillating fiber that may lead to nonuniform spatial sensitivity effects. 1. Geometric coupling efficiency 2. Reflective vs. nonreflective efficiency of the reflector 3. Reflection efficiency at the uncoupled end of the scintillating fiber 4. Attenuation of light at various positions in the scintillating fiber Number 1 considers the loss in spatial sensitivity from misalignment between the scintillator and the optical fiber inducing a geometric spatial efficiency effect. This effect arises if the diameter of the scintillator is larger than the diameter of the optical fiber, leading to escaped light photons. In addition, any misalignment between the two fibers during construction could potentially magnify this effect. Number 2 explains the effect of differences in sensitivity of the fiber at the surface of the optical fiber where light can be transmitted without being reflected within the scintillator compared to light that must be reflected. In this case, the transmission cone, defined by the critical angle of internal reflection within the scintillator, is greater than the acce ptance cone, defined by the numerical aperture of the scintillator waveguide surface. Number 3 accounts for the uncoupled end of the fiber reflecting less than the entire incident light. If the reflector was 100% efficient, then transmission losses would be negligible because light would travel the same distance when assuming isotropic light emission. Therefore, any reflector with less than 100% light reflection efficiency will be less sensitive to detecting light emitted on the reflector side of the fiber. Number 4 includes the length of fiber and the attenuation coefficient, and evaluates their effect on the spatial sensitivit y

PAGE 114

114 Plastic Scintillating Dosimeter s Plastic scintillating dosimeters (PSDs) have been used extensively for radiological dosimetry applications.78, 99, 121 Hyer examined the utility of PSDs in the CT energy range and estimated that measurement uncertainties for evaluating organ doses were no mor e than 4%. Contributing factors included energy dependence due to changes in beam quality and scatter, angular dependence from a full normal to axial irradiation in a scatter medium, reproducibility, and variation in response from bend radius. Most impor tantly, these dosimeter s operate in real time and because of their small size and flexibility can be easily be used with phantoms providing the user with accurate and rapid evaluation of organ doses. The fabrication of these dosimeter s for use in CT can be achieved by simply modifying the length of the scintillating to any desired length that would cover the organ or target volume desired. Materials and Methods Simulated Spatial S ensitivity Simulations were conducted to investigate the applicability of an elongated plastic scintillating fiber by evaluating any variation in spatial sensitivity. The plastic scintillating fiber and optical waveguide used in this study were commercially available products with physical and optical properties listed in Table 3 1 A simple calculation reveals that the transmission cone of the scintillating fiber (42.74) is greater than the acceptance cone of the optical fiber (15.81). This satisfies the third condition since unreflective light reaching the scintillator wave guide boundary will be transmitted to the optical fiber with the same efficiency as reflected light. Light propagation in fiber optics is often restricted to analysis of meridonial rays, despite both skew and meridional rays traveling within a fiber optic cable. This type of analysis was justified by Cozannet and

PAGE 115

115 Treheux122 provided the appropriate acceptance angle of skew rays is used. Meridoni al rays are less complex since the ray is confined to the plane containing the fibers optical axis and the point at which the ray originated. Skew rays travel in a nonplaner zig zag path that never crosses the axis of the fiber. Knowing the numerical aperture of the optical fiber, the core refractive index of the scintillating fiber, and optical fiber; the effective numerical aperture and effective acceptance c one of the optical fiber and scintillator is defined in Equation 41. = sin = ( 4 1 ) Equation 41 was referenced from Beddar et al .107, which was used to model the light collected from the acceptance cone of the scintillator optical fiber coupling. A MATLAB (Mathworks Inc., Natick, MA) script function was written to evaluate the spatial efficiency of the scintillator for different lengths and attenuation coefficients. Equation 4 2 formulates the fraction of light attenuated by the scintillator at different distances from the coupling surface. ( ) = 1 2 + 1 2 ( ) ( 4 2 ) photon attenuation coefficient. The integral is independent of radial dist ance from the central fiber axis because, although it will induce maximum one additional reflection, it does not change the total distance traveled within the scintillator. The exponential is integrated over all angles less than the acceptance angle defined in Equation 4 1. A separate script, modeling the light loss from fiber alignment and reflector efficiency, was coded to determine any axial spatial dependence from a uniform

PAGE 116

116 exposure. The method of evaluating the alignment effects involved calculating the success that a light ray emitted at a position within the phantom at a specific angle would cross the fiber boundary directed into the core of optical waveguide. This type of analysis is illustrated in Figure 41 T he coupling efficiency at a position ( x ) may be evaluated by integrating over the passable locations and the nonpassable locations designated by the white and dark areas respectively. The location the light ray crosses the surface is determined by its y position. The y position c an be determined by the angle it is emitted and the distance from the surface as seen in the schematic. Light contributions from the cladding material were ignored due to complexity. At each position (x) the probability (p) may be evaluated by integrating over p(y), which is expressed in discrete voxels which represent discrete radial positions (y). The chosen intervals were 0.005 mm. For simplification, a voxel along the axis was assigned a number 1 if it was within the white region or 0 if it was in the dark region. Construction The linear fiber optic dosimeter (LFOD) was constructed by modifying the sensitive element of the fiber optic coupled (FOC) dosimetry system. The plastic scintillating fiber (BCF 12, Saint Gobain Crystals, Nemours, France), which was previously 3 mm in length, was replaced to mirror the dimensions of organs for the University of Florida adult anthropomorphic phantom. Therefore, a series of scintillating fibers were fashioned to variable lengths representing the typical dimensions of an adult human organ: 3.0, 5.0, 7.0, 10.0, and 15.0 cm. One end of the scintillating fiber was coupled to a multi mode transmission optical fiber (400UV, Polymicro Technologies., Phoenix, AZ), having a 400Before coupling, both ends were polished to improve coupling efficiency. The remaining end of the

PAGE 117

117 scintillator was coated with reflective paint (EJ 510, Eljen Technology, Sweetwater, TX) to increase the acceptance efficiency of the fiber optic system. The coupled scintillating transmission fiber assembly is encased in heat shrink tubing, mounted to a SMA connecter (SMA 490, Ocean Optics Inc., Dunedin, FL) and sealed to prevent ambient light contamination. The area of heat shrink tubing covering the sc intillating fiber was spray painted to identify the sensitive element. A separate LFOD was constructed to include a reference fiber for dosimetry measurements. This was necessary to quantify the stem counts generated in the optical fiber from the natural florescence when exposure to ionizing radiation. In order t o calibrate and quantif y th e air kerma accurately the dosimeter was made with a 10cm scintillating fiber, the same active length as a 10cm pencil chamber ionization chamber. Characterization General c haracterization The characterization of the LFOD for spatial sensitivity, energy dependence, reproducibility, and linearity was performed using a clinical radiographic x ray tube (CPI Indico Plus, Communications & Power Industries International Inc., Georgetown, Ontario, Canada) with variable output and characterized half value lay er of 3.66 mm Al at 80 kVp. All measurements were repeated five times to reduce statistical uncertainties and provide reliable reproducibility measurements. A 100 m m pencil ionization chamber and associated electrometer (chamber model RC3CT, Accu Dose ele ctrometer model, RadCal Corporation, Monrovia, CA) were used to provide simultaneous exposure measurements. The linearity and reproducibility of the two LFOD s were evaluated simultaneously by delivering a variable exposure by increasing the tubecurrent product from 32 to 400

PAGE 118

118 mAs. The modification of the tubecurrent product was done by varying both the exposure time and tubecurrent. In effect this method also represented the dose rate linearity of the dosimeter. However, from the previous chapter the dose rate of the dosimeter was been well characterized. A reference chamber (pencil chamber with cap ) was used to quantify the reference air kerma for each tube setting. The detectors were placed on top of a backscattering medium (PMMA) at 40 cm from t he anode. The field size was collimated to 5cm x 5cm at the detector measurement location to ensure the x ray beam was contained by the sensitive element of the dosimeter and ion chamber. Energy dependence was evaluated by delivering xrays from a range of tube potentials at the LFOD and reference chamber positioned on the central axis free in air Several x ray tube voltages of 80, 90, 100 110 120 and 130 kV were selected for this study. The pencil chamber with the cap, and LFOD were centered along the cross hairs of the x ray tube and placed orthogonal to the axis that runs parallel to the variable x ray intensity to avoid the heel effect. The x ray field was collimated to a 7 cm by 7 cm field at the detection depth (40cm SDD) T he energy dependence was quantified by taking a ratio of the LFOD readout to the measured air kerma, expressed in counts/mGy, and plotting this value normalized to the 80 kV correction factor for each tube potential. For completeness in evaluating the dosi meter s response to changes to beam hardening varying thicknesses of soft tissue equivalent material were placed on top of both the dosimeter and pencil chamber with the cap, to evaluate effects from beam hardening and scatter. Two custom rectangular ti ssue equivalent phantom s were

PAGE 119

119 constructed to house the ion chamber and dosimeter and a set of rectangular slabs. Figure 42 illustrates the detector phantom with slabs of soft tissue equivalent material slabs placed on top. Exposures were delivered at 120 kV, 50 mAs, a nd a fixed field size of 5x5 cm. which was measured at the detector depth (60cm SDD) to ensure the full beam, comprised of scatter and primary x rays, would be contained by the sensitive volume of each detector Similar to the previous experiment, a plot was created containing the normalized detector sensitivity in counts/mGy, versus depth in the material. Spatial sensitivity was evaluated by sampling counts as a function of length exposed to a uniform x ray field. A step motor (Model 1000, Advanced Radiation Measurements, Port Saint Lucie, FL) was used to insert each LFOD incrementally into the x ray field generated by a clinical portable x ray unit at 80 kVp and 10.5 mAs. The LFODs were placed in the phantom with 1.5 cm of material to attenuation and backscatter. Sensitivity was measured in number of measured counts as a function of exposed scintillating fiber length (sensitive region) in 5 mm increments. Variable spatial intensity from the heel effec t was avoided by inserting the LFOD into the x ray field perpendicular to the axis containing the heel effect Characterization for c omputed t omography d osimetry T he effect s of spatial sensitivity in a phantom from a CT irradiation were evaluated. The 1 0 0 m m pencil chamber was used to quantity reference air kerma values from varying beam profiles delivered along the axis of rotation which was aligned with the center hole location of a CTDI body phantom Beam profiles were varied by adjusting the collimation width Exposures were delivered from a Toshiba Aquilion One volumetric scanner (Toshiba America Medical Systems Inc., Tustin, CA) in a single

PAGE 120

120 rotation with no table translation to each detector free in air with variable collimation width of the x ray beam along the axis of rotation while keeping the exposure parameters held constant. The maximum selected collimation width was 8 cm to include the fall off from scatter and attenuation through the bow tie filter The measurement was performed freein ai r because of the wide profiles associated with inphantom measurements attributed to scatter radiation.8 Boone demonstrated through computational methods that efficiency of measuring CTDIvol is quite low after 10mm slice acquisitions. Therefore the freein air measurement technique was selected as the optimal method of comparing air kerma values from the 10 0 m m pencil chamber and the LFOD A unique characterization method was required to characterize the LFOD for CT dosimetry to approximate the actual irradiat ion conditions experienced by the dosimeter. For CT dosimetry characterizations, a CTDI body phantom provided the most realistic physical representation of the bodys attenuation characteristics while als o able to accommodate an ionization chamber to quan tify reference air kerma for the 10 0 m m LFOD The effects of penetration depth, beam energy, scatter radiation, and spatial sensitivity were all evaluated simultaneously by quantifying the ratio of detected light photons to air kerma from different beam energies For this characterization a different reference chamber was selected. CTDI pencil chambers are acceptable cavity type chambers as long as the beam (primary and scatter tail) is confined within the sensitive volume of the detector. However, for m ost clinical scans the scanning region will fall outside of the 100mm chamber and the correlation between air kerma and dose to soft tissue will breakdown.

PAGE 121

121 To evaluate the dose within the phantom t he AAPM protocol for evaluation of CT dose was adopted. From this protocol the axial dose profile is saturated by performing a full scan with table translation using a narrow beam over an e longated CTDI phantom with a detector centered on the machine isocenter and placed at the midline of the phantom Figur e 42 illustrates the setup geometry and response of the LFOD. By using a narrow collimation width and a long phantom, t he integrated dose will approach an equilibrium value and generate a flat dose profile over the center hole location within the phantom This equilibrium dose is characteristic of the tube voltage, mAs, filter shape, and pitch. In this scenario, the tube voltage was varied to produce a energy dependent response at the detector position that includes a primary and scatter component. The phantom was assembled using a series of adjacent CTDI body phantoms and a pair of tissueequivalent cylindrical phantoms on each end. Measurements were taken for three tube voltages ( 100, 120, and 140 kV) at the center hole location. Equilibrium was established for LFOD when the integral counts at later time intervals approached zero at the beginning and end of the scan. Results Computational Simulations Each MatLab simulation was performed to vary the highlighted parameters that were deemed significant to affect the spatial sensitivity of the LFOD The default parameters for fiber length, attenuation, reflector efficiency, and alignment were taken from the manufacture information for the s components highlighted in Table 4 1: 10 cm, 0.00370 cm1, 0.95%, and completely aligned. Figure 4 4 shows the effects of the reflector efficiency and effects from fiber alignment. Figure 4 5 shows the effects of fiber length and attenuation. These simulations indicate a higher importance of fiber

PAGE 122

122 alignment, and a lesser importance of reflector efficiency. The fiber light attenuation coefficient plays an important role as well, especially at greater lengths. The specified attenuation coefficient by the vendor was 270 cm1 indicating that fiber alignment plays the mo st significant role in determining the spatial independence of these types of detector s. In general, this effect is most observable within the first 12 cm of the fiber from the coupling side of the scintillating fiber Linearity and R eproducibility Tab le 4 2 summarizes the average readings and standard deviation sampled from 5 measurements over the exposure range delivered from x ray t ube current time products from 32 to 400 mAs. The data was fit to a linear nothreshold function to determine the sens itivity of each fiber. At the bottom of Table 4 2 the linear correlation coefficient is displayed to evaluate the linearity of the data along with the sensitivity expressed in terms of counts per air kerma (mGy) Energy Dependence The 10 cm LFOD containing the reference fiber to quantify background radiation was used to match the length of sensitive element to the penci l ion chamber. The normalized readings, in counts/mGy, were plotted against the selected tube potential Th e data is displayed i n Figure 4 6 The response was normalized to the 120 kV data point The energy response from changes to tube voltage is shown to be positively linear, with an average 0.6 % increase per kV For reference, the PSD using a 3mm scintillator was characterized as a positive energy dependence of 0.68 % per kV offering a similar free in air energy response. The depth/penetration dependence of the LFOD is shown in Figure 4 7 Each reading was normalized to the surface (zero penetration depth) data point to illustrate

PAGE 123

123 the positive response with increasing tissue penetration. This positive increase varied 11% over the 14 cm of soft tissue equivalent material. The positive slope has been attributed to beam hardening which results in an x ray beam of higher averag e energy. At energies between 40 and 70 keV, the ratio of mass energy absorption coefficients between polystyrene and air is positively linear as well. The reported variation in the PSD was described as 10% across the 16 cm of material. Overall, the energy response of both dosimeters is very comparable indicating little importance on scintillator length and inaccuracies associated with beam energy. Spatial S ensitivity Spatial sensitivity was evaluated for the series of LFODs by evaluating the response to the length of scintillating fiber irradiated in the x ray field. The six scintillators, with sensitive element lengths of 3.0, 5.0, 7.0, 10.0, 12.0, and 15.0 centimeters were all evaluated. A step motor was used to insert incrementally into the x ray field axis perpendicular to the heel effect. The x ray field was generated by a clinical portable x ray unit at 80 kVp and 10.5 mAs An ion chamber verified the uniformity of the area of the field used for exposures to within 5%. Sensitivity was measured in number of measured counts as a function of exposed scintillating fiber length (sensitive region) at 0.5 cm intervals. A linear fit was applied in order to obtain the number of counts per centimeter of exposed fiber and along with the linear correlati on coefficient. The linearity of the LFOD s proved to be excellent demonstrating good uniform spatial sensitivity. The lowest linear correlation coefficient was calculated to be 0.9933. Each detectors response as a function of exposed length is plotted in Figure 47.

PAGE 124

124 Characterization for Computed Tomography Dosimetry Figure 49 shows the response of the ion chamber and LFOD to varying beam collimation widths irradiated freein air. The tube voltage was set to 120 kV, the tubecurrent to 200 mA, and the slice thickness was maintained to 0.5 mm. The beam width was changed by selecting variations of number of slices per rotation. For example, a 64 slice scanner, with a 0.5 mm slice thickness would use a 3.2cm beam. The beam width is measured along t he isocenter, which corresponds to the center hole location. Beam collimation thicknesses ranged from 0.5 mm to 7 mm in order to contain the entire beam within the sensitive length of each detector. The response by the LFOD was exceptional, again proving no spatial sensitivity effects. Scan parameters were as follows. The full scan was performed over 43 cm of acrylic and soft tissueequivalent material at a pitch of 1.5. The beam was collimated to 20 mm in order to perform the helical scan with a tube current of 200 mA, 0.8 s rotation t ime, and 30 mm/rot table translation. The tube voltage was selected from 100, 120, or 140 kV using the medium body filter. Data was captured using the detection system to record signal at each time interval (0.1s) in addition to the total number of count s detected. Three measurements were taken at each tube voltage and then combined together and averaged to generate three profiles. The time scales were aligned together to create a time index in order to account for the scans start ing at different times. Each scan was 12.9 seconds, therefore 129 measurements contributed to the profile. Figure 410 plots the profile from each tube voltage as well as a fit function for all three profile using a Gaussian distribution. The profile is actual a relative prof ile taken by normalizing the counts in each bin by the maximum for that energy. The Gaussian distribution is actually a two component Gaussian, one modeling the scatter

PAGE 125

125 radiation contribution, and the other for primary beam radiation. The two component G aussian distribution which includes the primary and scatter component is plotted in Figure 411A Irradiating over the entire scan length results in 50% of accumulated/equilibrium dose is from scatter, but at the center when the beam is directly irradiat ing the detector, only 30% is of the dose is from scatter. These implications may affect how detectors are calibrated moving forward in CT. To demonstrate th e effect of scatter, Table 4 3 shows the relative air kerma correction factors (normalized to 14 0 kV) at the center hole location of the CTDI phantom for the gadolinium scintillating detector (GSD) described in Chapter 3 and the correction factors for the LFOD The correction factor for the LFOD varies by approximately 0.3% per kV which is half of the freein air energy dependence of 0.6% per kV. The gadolinium detector performs extremely well, especially when compared to the LFOD, which can be attributed to the majority of scatter radiation that it receives compared to the primary radiation. This is also illustrated in Figure 4 11B which shows the broadness and magnitude of the gadolinium scatter profile compared to the primary curve Scatter radiation, or low energy radiation, has very similar energy absorption characteristics among materials because x ray interactions are dominated by photoelectric absorption which results in all of the x ray energy absorbed locally. T he energy dependence of the LFOD was co rrectable by using the scatter and primary profile as well as the measured air kerma to generate a primary and scatter correction factor. The count data indexed at each time step was binned into scatter radiation and primary radiation based on the Gaussia n distribution in Figure 411 A The counts in each bin were summed to get net scatter counts and net primary counts.

PAGE 126

126 Using the measured air kerma from the ionization chamber, a MatLab program was constructed to find the optimal values of the scatter and primary correction factors. The primary correction factor was energy dependent and the scatter coefficient was the same for each producing four correction factors. These correction factors are listed in Table 43 in the far right column. Using these correction factors and the LFOD data, the air kerma could be estimated to less than 1%. Discussion The LFOD demonstrates very sim ilar characteristics to the PSD This includes an exceptional linear response with beam intensity and a positive linear energy dependence within the exposure range of radiographic and CT imaging modalities The overall energy dependence indicates an increasing linear trend with respect to x ray tube with an estimated variation in sensitivity of 0.6 % per kV. Good spatial sensiti vity was observed for each of the LFODs used The spatial sensitivity did appear to regress with increased length, which was an expected result from the simulations. The geometric coupling effect was observed for a few of the dosimeters and is indicated at the most distal lengths of the spatial sensitivity evaluation since the scintillator optical fiber interface were inserted last in the experiment. From the freein air characterization, the dosimeters showed no loss in sensitivity with respect to the ion chamber. One of the benefits of CT is that the dynamic motion of the gantry tends to reduce the effects of penetration depth leading to a decreased variation in sensitivity. This was demonstrated by Fisher who used a 32cm CTDI bo dy phantom to calibrate an FOC for CT organ dose measurements.123 This was further validated by showing that scatter radiation minimizes the energy dependence of a detector. However the LFOD is

PAGE 127

127 less prone to scatter than a point detector because its sensitive element spans centimeters in length. Therefore, it shows higher energy dependence than a point detector. However, it was demonstrated that along the central axis, that the count profile can be used to correct the dosimeter for energy dependence from the primary x ray beam by fitting the profile to a dual component Gaussian distribution and calculating the appropriate correction factors. The main advantage of this type of detector is that its sensitive length can be used for organ dosimetry by using a scintillating fiber similar to the length of the target organ. This detector will be especially useful in areas of scan termination and initiation as well situations where the surrounding tissue is varying creating different intensitys along the imaging axis. Most co mmercial phantoms allow for dose locations in the x y plane and are separated in the z plane in some cases up to 1020 cm. Anatomy changes as well as tube current modulation will lead to dynamic beam profile that cannot be captured using point dosimeters and would be more accurately evaluated using integrated detectors. Conclusion It was d emonstrated that the LFOD exhibit s the desirable characteristics of a dosimeter that would be very useful for CT dosimetry. Similar to the PSD, its main drawback is th e energy/depth dependence from changes to the x ray energy spectrum. The LFOD from a general characterization standpoint mirrors the PDSs response. Both shared an exceptional linear response to beam intensity and a positive free in air energy dependence. T he LFOD detector geometry allows for a lower variation in detected counts, due to the higher light output from the scintillating fiber. The energy dependence is minimized by using the dosimeter in a high scatter environment versus

PAGE 128

128 free in air. Overal l, this type of detector appears most appropriate for measuring organ doses, where the scintillating element can be fashioned to the dimensions of the organ being evaluated. The rotating x ray tube will minimize the angular dependence of the fiber as well as flatten the depth dose profile. In addition to providing rapid and accurate measurements, when properly calibrated, this type of dosimeter demonstrates exceptional utility for in phantom CT organ dose measurements.

PAGE 129

129 Table 4 1. Optical p rop erties of the plastic scintillator and optical waveguide Plastic scintillating fiber (BCF 12) Silicon optical fiber (FVP400440480) Cladding Cladding material: Acrylic material: Silica refractive index: 1.42 refractive index: 1.46 thickness: thickness: Core Core material: Polystyrene material: Silica refractive index: 1.6 refractive index: 1.44 diameter: diameter: Numerical aperture: 0.58 Numerical aperture: 0.22 Attenuation coefficient: 2.70 m Attenuation coefficient: 8.33 m

PAGE 130

130 Table 4 3. Beam profile data from the full body phantom scan at 100, 120, and 140 kV. Tube Voltage Relative air kerma c orrection factor Total scatter fraction* Scatter f raction along phantom central axis* LFOD c orrection f actors (10000 counts/mGy) (kV) GSD LFOD GSD LFOD GSD LFOD Primary Scatter 100 0.990 0.875 0.832 0.507 0.416 0.301 2 72 0.98 120 1.002 0.942 0.832 0.507 0.416 0.301 3.57 0.98 140 1.000 1.000 0.832 0.507 0.416 0.301 4 .8 3 0.98 Taken from dual component Gaussian fit where the broad curve represented the scatter contribution and narrow primary. Table 4 2 Reproducibility and linearity measurements with 7 cm and 15 cm linear dosimeter s Tissue kerma 7 cm FOC linear dosimeter 15 cm FOC linear dosimeter (mGy) Mean counts Standard deviation COV (%) Mean counts Standard deviation COV (%) 1.5 15289 165.12 1.08 110062 236.17 0.215 1.87 18916 236.45 1.25 135945 487.68 0.359 2.93 29578 300.22 1.015 210977 490.81 0.233 3.73 37333 176.96 0.474 264227 1090.82 0.413 4.68 46830 220.57 0.471 323149 425.21 0.132 5.85 58691 117.38 0.2 395438 212.08 0.054 7.5 75156 95.44 0.127 506237 256.19 0.051 18.77 189022 489.89 0.256 1272686 1958.27 0.154 Sensitivity: 10059 ( counts/mGy ) Sensitivity: 68029 ( counts/mGy ) Corr. Coeff: 0.99999 Corr. Coeff: 0.99955

PAGE 131

131 Figure 41. Illustration of the coupling efficiency analysis. At a given (x, y) location within the fiber, lines can be drawn from angles within the acceptance cone to the coupling surface which would result in light rays be transmitted or lost due to the uncoupled areas on the scintillator. The variable x is the length along the fiber which gets emitted at an angle theta. This light rays would fall in a voxel (black or white). The voxel would then shift based on the y position. dy d x y 1 reflection Probability (100% or 0% for different voxels (white or dark) x

PAGE 132

132 Figure 42. Tissue equivalent phantoms for depth dose measurements. The left (a) is the LFOD phantom while the right (b) is the pencil chambers housing. Both phantoms are placed onto of acrylic back scattering medium. Photo courtesy of Matthew R. Hoer ner a) b) a) b)

PAGE 133

133 Figure 43 Phantom assembly and centered LFOD moving together from the right to left over the scanning l ength. The plot on the right shows the accumulated dose as a function length scanned. As the scan length increases the integrated dose will approach an equilibrium value. The scan was repeated for the ionization chamber and the equilibrium dose was meas ured. Scanner axis of rotation Scanner axis of rotation Scanner axis of rotation Scanner axis of rotation Scanner axis of rotation Length Dose Length Dose Length Dose Length Dose Length Dose

PAGE 134

134 Figure 44 Relative transmission of light through the optical scintillating fiber boundary at locations along the scintillating fibers axis. The simulation was run to vary the attenuation coefficient of light within the detector ( ) and the length of the fiber. All other parameters were set to the values listed in Table 4 1. 0 2 4 6 8 10 12 14 16 18 20 0.95 0.96 0.97 0.98 0.99 1 1.01 Distance from coupling surface (cm) Relative transmission 10-cm fiber, = 270 cm-1 10-cm fiber, = 70 cm-1 20-cm fiber, = 270 cm-1 20-cm fiber, = 70 cm-1

PAGE 135

135 Figure 45 Relative transmission of light through the optical scintillating fiber boundary at locations along the scintillating fibers axis. The simulation was run to vary the efficiency of the reflector ( ) and alignment of the optical fiber and scintillating fiber. All other parameters were set to the values listed in Table 4 1. 0 1 2 3 4 5 6 7 8 9 10 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1 1.01 Distance from coupling surface (cm)Relative transmission = 0.5, 0.1cm misaligned = 0.95, 0.1 cm misaligned = 0.5, aligned = 0.95, aligned

PAGE 136

136 Figure 46. LFOD response factor measured for varying x ray tube voltage (80 through 130 kV). Measurements taken freein air using a 100mm pencil chamber to quantify reference air kerma. Error bars correspond to 1 standard deviation of the mean. Figure 47. LFOD response factor measured for varying penetration depths into a soft tissue equivalent medium (0 through 14 cm). Detector position was fixed with 5 cm of backscattering material. Error bars correspond to 1 standard deviation of the mean. 70 80 90 100 110 120 130 140 0.9 1 1.1 1.2 1.3 Tube voltage (kV) Normalized counts / mGy 0 2 4 6 8 10 12 14 16 1 1.05 1.1 1.15 1.2 1.25 Depth (cm)Normalized counts \ mGy

PAGE 137

137 Figu re 4 8 Evaluation of the 7cm, 12cm, 10 cm, and 15cm LFOD dosim eters for spatial sensitivity using a uniform intensity x ray field. Linear fit was applied to the each data set along with a linear correlation coefficient. Error bars correspond to 1 s tandard deviation of the mean. 0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16 x 104 Irradiated length (cm)Counts (ten thousands) 7-cm dosimeter 12-cm dosimeter 15-cm dosimeter 10-cm dosimeter

PAGE 138

138 Figure 49. LFOD response to varying collimation widths. LFOD response was benchmarked by measuring the reference air kerma using a 100mm pencil ion chamber (mGy). Residual plot below shows the relative error of the linear nothreshold fit to the data. Error bars are shown in to show the +/ 1 standard deviation from the mean. 0 1 2 3 4 5 6 0 2 4 6 8 10 x 104 Air Kerma (mGy) LFOD response (counts) 0 1 2 3 4 5 6 0.95 1 1.05 Residuals

PAGE 139

139 Figure 410. Distribution of counts from discrete time intervals sampled over a helical CT irradiation of a 10cm LFOD in a custom CTDI phantom. Three of the curves correspond to each of the tube voltage settings and the black line is the two component Gaussian fit. 0 20 40 60 80 100 120 140 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Time indexNormalized counts per sampling frequency 140 keV 120 keV 100 keV fit

PAGE 140

140 Figure 411. Gaussian fits of each counting profile from ( a) LFOD and ( b) GSD. 0 20 40 60 80 100 120 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Time indexDistribution (normalized to maximuma) total primary beam scatter radiation 0 20 40 60 80 100 120 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Time indexDistribution (normalized to maximumb) total primary beam scatter radiation

PAGE 141

141 CHAPTER 5 CONTRUCTION OF AN ANTHROPOMORPHIC PREGNANT FEMALE PHANTOM SERIES FOR COMPUTED TOMOGRAPHY DOSIMETRY Physical Phantoms for Experimental Radiation Dosimetry Anthropomorphic phantoms used for dosimetry applications have been an accepted method for measuring organ and tissue doses from medical radiation procedures.121, 124126 The advantage of using this tool for patient dosimetry is the anatomy of human subjects is not required. Anthropomorphic phantoms are used in dosimetry studies primarily to: simulate the human bodys absorption and scatter of ionizing ra diation, incorporate dosimeter s throughout their volume, and correctly model the position, size, and orientation of the organs of interest residing within the phantom .127 In general, they offer varying degrees of complexity with the most complex phantoms designed for a specific imaging modality or representation of a specific anthropometric population. Phantoms built specific ally for CT dosimetry are typically constructed in cross sections to accommodate placement of the r moluminescent and optically stimulated dosimeters (TLDs and OSLDs) and to avoid adjoining phantom components in the imaging plane to create a uniform and realistic image The tissue equivalent materials are characterized to have matching density as human tissue. Custom built anthropomorphic physical phantoms can be a beneficial alternative to the prohibitive cost of the commercially available CT phantoms and address the specific demands required to produce accurate dosimetry One of the benefits of building anthropomorphic phantoms is that they can be customized so that dosimeter s could be placed along the z axis ( cranial caudal axis ). In the previous chapter, it was demonstrated that linear fiber optic dosimeters (LFODs) are a good dosimeter for

PAGE 142

142 evaluating dose to large volumes that would otherwise require mult iple detectors. Their main advantage is that they can be positioned so that their sensitive length aligns with the z axis spanning the width of the internal anatomy. This type of detector orientation is not possible using commercially available CT phanto ms. University of Florida Physical and Computational Phantom Series A series of anatomical reference physical phantoms have been fabricated for research applications at the University of Florida (UF).128 The physical phantoms at UF are built slice by slice using three classifications of tissue equivalent substitutes representing the three types of bodily tissue most differentiated by their attenuation characteristics: lung tissue equivalent substitute (LTE S), bone tissue equivalent substitute (BTES), and soft tissue soft tissue equivalent substitute (STES), which like other phantoms is a composite material representing the remaining anatomy The anatomy in each cross section is derived from its virtual blu eprint, a computation hybrid phantom provided by Dr. Wesley Bolchs research group at UF .129131 A computational hybrid phantom is defi ned as a voxelized phantom that has been stylized to meet reference criteria which includes: height, weight, and organ mass and volume.132 The shape and location of anatomy is not specified for a reference phantom, but hybrid phantoms maintain a level of accurately since a hybrid phantom uses actual human anatomy providing more real i sm Building physical phantoms following the sliceby slice method has shown to be an exhaustive process taking months to accomplish with extra attention required to each slice from sanding of t he BTES, trimming of excess STES material, and engraving time. Previous phantoms required anywhere between 10 to 20 minutes of engraver time for creating a single soft tissue mold.

PAGE 143

143 Research Based Pregnant Female Anthropomorphic Phantoms Currently there exist commercially available phantoms that anatomically model a reference women or reference male/female for abdominal and pelvis studies.17, 19 133 A few have used physical phantoms to perform dosimetry on unborn children in the first trimester. In one study, the abdomen/pelvis region was enlarged to portray the extra weight added over the course of pregnancy.20 The phantoms utilized by these studies are listed in Table 51 with their associated cost Using an adult reference female phantom will limit measurements to obtaining organ doses not fetal doses In the beginning stages of pregnancy this is sufficient. However as the pregnancy progresses and the fetus begins to fully develop it becomes uncontained by the volume of a nonpregnant female uterus and along with the expansion of mother s waist and pelvis the modeling of anatomy of a reference female becomes insufficient Studies have shown that fetal dose strong ly correlates with fetal depth and diameter of the pelvis making pregnant female physical phantoms modeling the later stages in pregnancy more valuable.21 Other s tudies have also demonstrated the effect of patient size on organ doses.134 135 Building a Novel Anthropomorphic Phantom A new construction method was created aimed to reduce the amount of labor exerted when constructing phantoms slice by slice while simultaneously providing CT dosimetry along the cranial caudal axis of the phantom In addition, using a higher z axis resolution provides more realistic anatomy, which was previously limited to 5 mm. The method utilizes constructing the phantom in components with the surfaces corresponding to dose locations relevant for obtaining dose information. Each section is constructed from a mold generated by a 3D printer By building the phantoms in volume

PAGE 144

144 sections less time is dedicated to treatment of indiv idual slices, and allows for mass production and introduction of tissue equivalent materials. Although the dosimetry is limited to points fabricated by the phantom with careful planning accurate dosimetry can be accomplished. With the development of a reference pregnant female series, the computational phantoms can be rendered rapidly for use in experimental dosimetry. Materials and Methods A series of tissue e quivalent substitutes were created to match the radiological parameters of tissues in the energy of diagnostics x rays. These tissue substitutes include: adipose tissue (ATES), glandular tissue (GTES), and a new soft tissueequivalent substitute (STES). All material substitutes were evaluated for three measurable parameters that best describe the low energy x ray properties of tissues: density, half value layer for CT beam spectra, and Hounsfield units. The techniques for evaluating each of these propert ies are described in detail below. F ormalisms for calculating the variance or uncertainty of using these materials are derived to be used in sensitivity studies. Sensitivity studies are performed using Monte Carlo techniques to evaluate the effects of material densities on the accuracy of obtaining doses. Typically a parameter such as density is varied and simulations are run to determine sensitivity of doses to material density. Density is a good parameter because it correlates linearly with the attenuation and absorption of x rays. Measurement Techniques Hydrostatic weighting technique The density for each material was evaluated using hydrostatic weighting, a method used for measuring the mass density of a human body. The advantages of using this method are volume information of the material (body) is not required and its

PAGE 145

145 real world application of measuring body composition. The procedure is derived from Archimedes principle using three measured values: the weight of the body in air, the weight of the body completely immersed body, and the density of the auxiliary liquid. The instrumen tation required to precisely measure the density is a high sensitive scale and an associated density kit (Mettler Toldeo) The density of the body can be calculated using Equation 5 1 = ( ) + ( 5 1 ) The density of a solid is determined with the aid with the aid of a liquid whose within the auxiliary liquid (B). The second measurement corresponds to the buoyancy force acting on the solid. pressure. The scale is capable of measuring within 0.0001 grams. The density of air is a function of temperature and pressure, and the density of the water is also dependent on temperature. The temperature of the liquid water was measured using a thermometer, and using a look up table for distilled water density tabulated with water temperature, the density could be determined. Using error propagation the associated error of this method is: = + + + ( 5 2 ) Taking the derivative of each term from Equation 12 and substituting into Equation 13 gives: = ( ) + ( ) + ( ) + 1 ( 5 3 )

PAGE 146

146 The measurable errors in both A and B coincide with the scales resolution which equals 0.0001grams. For the measurement in water 0.000 8 was used as a conservative estimate. The error in liquid density is 0.0003 and the error in air density is about 0.00003. The fluctuation in density from pouring of the material was also evaluated by measuring the density of the material at different locations and then added to the error in using the density kit to obtain a final value for the variation in density: = + ( 5 4 ) Hounsfield n umber The Hounsfield scale is an attenuation scale relative to water that associates a number to each voxel in an image acquired from a CT slice based on the relative attenuation of the material within that voxel. The CT number, or Hounsfield unit, is defined using Equation 5 5 voxel and linear attenuation H20 = 1000 ( 5 5 ) The Hounsfield units (HU) were evaluated using a Toshiba A quilion One volumetric CT scanner (Toshiba America Medical Systems Inc., Tustin, CA) at standard imaging parameters with tube current modul ated exposure control. The average HU was determined from the selected regions of interest using clusters of pixels with areas approximately 10 cm2. Linear a ttenuation c oefficient The linear attenuation coefficient is a quantity that characterizes how easily a material can be penetrated by an x ray beam. This quantity is dependent on the

PAGE 147

147 incident beam energy or spectra and the physical properties of the material X ray or photon attenuation is modeled as an exponential function, expressed below in Equation 5 6. The intensity, X, transmitted by a thickness t from X0 the initial intensity incident to the attenuating medium. = ( 5 6 ) The linear attenuation coefficient was measured experimentally using a clinical CT scanner to provide a realistic CT energy spectrum. The beam was collimated into a narrow beam geometry using a custom lead plate to reduce scatter contribution. Several samples were fashioned into 1 cm thick slabs to be used for evaluating the linear attenuation coefficient. Each samples thickness was measured using a digital caliper (General, New York City, NY). The samples were then placed consecutively in the x ray beam and exposure measurements were taken after the introduction of a new sample. A curve was generated plotting exposure in logarithmic scale as a function of material thickness. The attenuation coefficient was determined to be the slope of the line. T he overall error from the attenuation coefficient determination is calculated using error propagation in Equation 56. = 1 + 1 ( + ) ( 5 7 ) The variance is calculated by rearranging the exponential attenuation, Equation 5 t is the thickness of the slab, and R is the natural log of the ratio of the unattenuated exposure, X0, to the attenuated exposure X. Additional sources of error such as the scatter contribution and

PAGE 148

148 beam geometry were not included as they would be reflected in the coefficient of determination from the data. In order to compare the measured attenuation coefficient to the theoretical, a target half value layer was calculated for the scanner s beam spectra. The target HVL was evaluated using Monte Carlo techniques. The source term for the clinical CT scanner, developed by Long and the ICRU 44 defined soft tissue, breast, and adipose tissue were simulated to determine a target HVL.136 The HVL was measured replicating the setup geometry of the experimentally measured HVL. The target HVL provided a necessary benchmark as the HVL is highly dependent on the x ray tubes inherent filtration which has shown to v ary scanner to scanner. For tissue substitutes that involve the STES material and polystyrene micro beads the linear attenuation coefficient may be calculated by using the following expression. = + ( 5 8 ) Since the ATES and LTES mixture is a composition of STES and polystyrene micro beads, the attenuation coefficient can be calculated from the mass fractions (w) of each component and the desired density of the adipose tissue. The value for poly was evaluated using the effective energy of the CT beam and the linear attenuation coefficients in the NIST database.108 In Attix, the weighting factors are expressed as (1g)w, where g is the radiation yield.84 However, for CT x ray energies, the value of g is essentially zero for the STES and polystyrene so the equation was reduced to just the weight fraction.

PAGE 149

149 Materials Soft tissue equivalent substitute ( STES) A similar urethanebased STES was designed to eliminate the density fluctuations in the previous urethanebased compound; PMC 121/30 Dry (SmoothOn, Easton, PA) combined with the CaCO3 powder ( Fisher Scientific, Hanover Park, IL), 2.8% by weight. A new urethanebased resin VytaFlex 40 (Smoothon, Easton, PA) was used in conjunction with URE FIL 7 (Smoothon, Easton, PA), a high density powder that dissolves well with urethane resins and improves overall stability. The powder also increases the overall effective Z of the material to better replicate the attenuation characteristics of soft tissue. The target density for human soft tissue is 1.04 g/cc. The accepted range of HU numbers for so ft tissue varies between 10 and 40. The STES was fabricated by combining equal amounts of the VytaFlex 40 constituents, and adding the appropriate amount of URE FIL 7 until the target density was achieved. An electric mixer was used to ensure a humongous mixture of the VytaFlex compounds and a uniform distribution of dissolved filler. The STES is allotted over 16 hours to cure and harden before measuring. After the target density was achieved, the compound was evaluated for attenuation and HU characteris tics. Several slices of STES material, approximately 0.9 cm thickness, were created to measure the linear attenuation coefficient for a CT x ray beam. Adipose tissueequivalent substitute ( ATES ) Although this material was never unintended to be incorpor ated into the pregnant female phantom, a series of tissue composites were made to represent bodily tissues similar to soft tissue but lower in density. The ATES was derived from the STES, described above, and adding poly fil polystyrene micro beads (Fairf ield Processing,

PAGE 150

150 Danbury, CT) to reduce the density. The STES density was modified to achieve a variety of densities corresponding to different breast tissue compositions and adipose tissue densities in the body. The homogenous breast and adipose tissues were modeled as the STES and adjusted to the desired density using the micro beads. A density curve was developed by adding various amounts of micro beads, by mass percent, and measuring the resulting density. Selective samples were further evaluated fo r their linear attenuation and HU properties using the techniques described in the above section. The development of these tissueequivalent substitutes only adds more detail to the phantom and providing a more realistic physical model for obtaining accur ate organ dose measurements. Bone tissue equivalent substitute ( BTES ) The BTES is an epoxy resin based material previously developed at UF The material is a mixture of a fiber glass resin and Calcium and Silicon powders to obtain a similar chemical structure as human bone. The composite is formed 51.0% by mass of the fiber glass resin (Bondo, 3M, St. Paul, MN) as well as 25.5% Silicon dioxide and 23.5% Calcium carbonate (Fisher Scientific, Hanover Park, IL). The BTES was developed as a homogenous bone mixture incorporating the cortical and trabecular spongiosa components seen in the human skeleton. The material was designed to match th e density, mass attenuation coefficient, and mass energy absorption coefficients to Oak Ridge National Laboratory (ORNL) defined values within the diagnostic energy range. Accepted HU numbers for bone tissue show a wide range of values from 400 to 1000.

PAGE 151

151 Phantom Construction Methodology Design and development of a natomy The pregnant female phantom series is constructed by Maynard from a set of computational phantoms derived from image sets of pregnant females who had undergone CT exams .137 Pregnant female anatomy was obtained from two sources of CT image sets: the Picture Archiving and Communication System (PACS) archive at the University of Florida, and publically anonymized images provided by Angel et al.21 All image sets were screened for coverage of the entire fetal area and no apparent physical abnormalities. Using the robust and flexible nature of NURBS/polygon mesh the contoured image sets were systemically combined with the UF adult nonpregnant computational phantom and four selected (2 week, 15 week, 25 week, and 38 week) UF reference fetal computational series. The 2 week and 15 week phantoms are very similar and can be combined to create a hybrid phantom segmented for the ovaries and uterus. These three phantoms were chosen from the eight phantoms in the library because they represented the significant gestation periods for peak radiobiological sensitivity for evaluating deterministic effects and cover the range of pregnancy for evaluating stochastic effects. T he design criteria for the phantom library was adapted to generate reference phantoms with: 1. A single, reference orientation within the gravid uterus 2. Age dependent target masses for the uterine wall, placenta, breast tissue, and total body mass 3. Preservation of original nonpregnant female reference tissue masses where appropriate, i.e. tissue not listed above 4. A single, reference maternal circumference, skin contour at each gestati onal stage

PAGE 152

152 For completeness, a reference fetal orientation was desired to be maintained throughout the phantom series. The left occipital anterior (LOA) configuration was adopted as the reference fetal position as it is the ideal at term configuration amo ng the possible vertex presentations, which account for the vast majority of fetal presentations during birth.138, 139 Target organ, tissue, and whole body masses were derived from ICRP publication 89 and used as the basis for adjusting or finetuning organ size and sometimes location to accommodate the fetus and match target reference mass data, as well as fine adjustment of external anatomy. The pregnant female computati onal phantoms are illustrated in Figure 51. Selection of f etal a natomy for d osimetry Each phantom was built from a base pha ntom which contained the anterior anatomy and the pelvic bone and lumbar spine. The base phantom was derived from the 25week phantom, containing no pertinent anatomy for dosimetry. The section of anatomy was further divided into components that each contributed a surface useful for dosimetry. A s a g itt al cut was done on each phantom that transverses the midline of the fetus. This cut split the fetus in half down the middle and provided the most area for dosimetry. Additional cut s relating to the head and body were performed in the axial plane to generate more dose locations for the 25 and 38 week phantoms. The number of cutting planes was limited to reduce the number of pieces and make 3D printing less complex The selected cut ting planes for each phantom are illustrated in Figure 5 2 The ovaries were selected to be segmented onto the 2/15week phantom to evaluate the dose for the embryo. This is visualized in Figure 52a which includes an additional cutting plane to include the ovaries.

PAGE 153

153 Fabrication using a 3D printer The computational phantoms contain identification tags to label the regions of each organ and tissue when voxelized to a binary format. A software program uploaded each phantom and isolated the soft tissue and bone region of the phantom with those specific tags and convert ed each volume to a stereolithography file (.stl) which is the file format for 3D printing. This process had to be done because the original bone is made of complex polygon mesh surfaces that cannot be directly converted into .stl file format After the soft tissue and bone structures of the abdomen/pelvic region were created they were modified to create a mold for casting. This was done using a robust computer aided des ign program ( Rhinocer ous v5.0 Robert McNeel & Associates, Miami, FL ) to extrude the surfaces into a new structure that encompassing the original surface The objective was to create a base phantom containing the lumbar spine and pelvis and soft tissue not useful for fetal dosimetry containing no vital fetal anatomy, but the mothers bone and muscle. The base phantom would provide structural support for the phantom when it would lay supine for dosimetry measurements during scanning. The first step was to render the pelvis and lumbar spine. The bone structure was divided into six molds: the lumbar spine, the sacrum, the left and right ilium, and the left and right ischium with the upper part of the femur. Each cast was built using a 3D printer containing the vacated cavit y to be filled with the bone substitute. The bone substitute was introduced to each mold using heat and to decrease viscosity allowing the material to settle in the mold. Long thin rods were used to compact the material into small cavities and help the m aterial settle. The bone substitute was allowed to cure over 24 hours, afterward it would be fully solidified. The open faces of the bone

PAGE 154

154 structure were sanded, and then brushed with water to soften the mold. Using a screwdriver, the mold began to crumble when applied with pressure. The six bone equivalent structures were then reassembled using a strong adhesive (Krazy Glue, Columbus, OH ) to form the complete pelvis and spine. The soft tissue component of the base phantom was built using 3D printer mol ds taken from the surrounding tissue of the pelvis and soft tissue posterior to the pelvis. The remaining sof t tissue was fabricated using similar method. First the anatomy was sectioned based on the dose locations. Then each piece was further divided t o meet printing specifications. Results Materials and Measurements The STES density was optimized by combining 9.0% by weight of the URE_FIL 7 powder and 91.0 % Vytaflex 40, which resulted in a density of 1.041 g/cc. The STES was further evaluated for density fluctuations within a single pour by separating the sample into six individual samples. The results showed that the density varied among ten samples by 0.0026 g/cc. The total variation in density, calculated using Equation 54, was 0.00261 g/cc. The linear attenuation coefficient was calculated to be 0. 228 cm 1 using linear regression. Applying Equation 18, the standard deviation in the measure d attenuation coefficient is 0. 0044 cm 1. The theoretical HVL calculation for the CT energy spectrum using the ICRU four component soft tissue showed an effectiv e attenuation coefficient of 0. 229 cm 1 of soft tissue. The HU values for soft tissue were between 0 and 20 averaging around 10. The ATES is the STES modified for density by adding the polyst yrene microbeads to offset the density. Figure 5 4 displays the density function of the ATES vs. the amount of added micro beads. Error bars were calculated using Equation 54.

PAGE 155

155 Fluctuations in density within a single sample were measured up to 0.003 5 g/ cc. ATES samples with densities of 0.98, 0.96, and 0.93 were selected to be appropriate for these bodily tissues: 50% (glandular) / 50% (fat) homogenous breast, 30% (glandular) /70% (fat) homogenous breast, and subcutaneous adipose tissue respectively. F rom Equation 58 using the weight fractions from the fit to Figure 54 and the attenuation coefficient of the STES the attenuation coefficients were calculated. The effective mass attenuation coefficient of polystyrene was determined by using the effectiv e energy of the CT spectrum. The homogenized bone density was measured to be 1.60 g/cc with a variation of 0.05 g/cc The half value layer of the bone material was measured to be 1.7 mm at 120 kVp, and when simulated using the density of the BTES (1.60 g/cc) was determined to be 1.61 mm From ICRU 46, the sacrum and lumbar column skeleton was listed to fall between 1.29 and 1.33 g/cc. This type of material is a homogenous bone tissue composition of the cortical and trabecular bone components. For comparison to other phantoms, the density of the Atom phantom is 1.60 g/cc and UF adult is 1.40 g/cc. The HU values fell between 650 and 810 with an average of 725. Completed Set of Phantoms All three pelvic/abdominal phantoms ( 2/ 1 5 hybrid, 25, and 35 week) are depicted in Figure 5 5 Figure 5 6 shows a to p ogram of the 1 5 week phantom. The phantoms were susceptible to air gaps where the pieces merged, and this was relatively unavoidable. However, it is easy to assemble and integrate with detectors. The color variation was due to using two different containers of one of the STES materials which varies in color among batches. The density did not fluctuate when switching containers. The mold was difficult to remove from the STES but relat ively easy from the BTES. A

PAGE 156

156 releasing agent (Smo oth On, Easton, PA) was used to aid the removing process but didnt fully remedy the situation. Discussion The new STES soft tissue has the c a pa bility to be modified for reaching the target density by controlling the amount of URE_FIL 7 powder introduced to the mixture. The hydrostatic weighting technique is a proven methodology of verifying the target density of the tissue substitutes. As phant oms are being constructed samples should be collected and measured for density to maintain consistency. This was deemed important because different volumes of tissue being fabricated lead to variations in the density therefore the formula had to be modifi ed to accommodate these deviations. Typically, this wasnt an issue in the old construction process because phantoms were built in slices. However, now that the phantoms were able to be built in a manner that allows for the mass production of materials, a quality assurance protocol was necessary With regard to the BTES it may be worth exploring a less viscous composite to ease the introduction of the substitute into the molds. Generating molds from a 3D printer proved to be an accurate and autom ated method of fabricating phantoms. However, mold materials that were unsmooth resulted in difficulty in extracting the STES material from the mold. The BTES was easy to remove from the mold. The bone molds should try to incorporate as many open faces as possible to ease the material introduction process due to the tight cavities formed within the bone mold. The bone material is viscous and doesnt pour well and has trouble settl ing into tight cavities without the aid of gravity or direct insertion. Theref ore consideration and caution should be exercised when determining the size of each mold. Mass production of tissue components were a luxury in expediting the construction

PAGE 157

157 process significantly. Pouring the entire soft tissue of all phantoms took man hours which is a significant improvement over the older method. In this situation the construction process is limited by how quickly the molds can be produced. It was determined that making bone molds this way is too difficult. The soft tissue mold should be fabricated to surround the bone material by creating voids in the mold to be filled later. Most printers currently use plastic materials which have similar physical properties to human tissue. It may be beneficial to use the plastic as casing to as semble the mold and limit the air gaps. This was a big issue with the pregnant female phantoms. Mistakes to avoid in the future are to cut the phantom into components with orthogonal angles, use a more release friendly material, and build the soft tissue and bone mold as a single mold. With the flexibility of modifying the density of the BTES and STES, a twocomponent BTES could be created by increasing the density of the STES to repre sent the marrow cavities of human bone, and the BTES represent the co rtical bone component. Conclusion This chapter explains a systematic approach of manufacturing computational phantoms for experimental dosimetry. This method has shown to streamline the production process of building phantoms by generating a small er number of molds that allow for the mass production of tissue substitutes while offering a higher resolution of anatomy and provide customized dosimetry locations In addition, the pregnant female physical phantoms are easy to assemble, produce high quality images, and are ideal for performing rapid dosimetry. The materials have been characterized to offer a range of tissue substitute options for incorporating more realism into each phantom.

PAGE 158

158 Table 5 1. Commercially available anthropomorphic phanto ms used for research Vendor Name Model Genders Cross Sectional Cost Supertech 3D sectional torso phantom CIRS 600 M Yes $17,664 Supertech Virtually human male pelvis phantom* CIRS 801 P F M No $19,156 The Phantom Laboratory RANDO phantom RAN100 M or F Yes $27,442 CIRS ATOM Dosimetry Verification Phantoms 702D M or F Yes $26,790 organs includes: bone, cartilage, spinal cord, vertebral disks, muscles, intestines, bladder, prostate, rectum, and interstitial fat ** all prices do not include plug ins for dosimeters

PAGE 159

159 a) b) c) Figure 51. Frontal view of the pregnant computational hybrid phantoms pelvis From left to right, a) 15, b) 25, and c) 38 week computational hybrid phantoms.

PAGE 160

160 a) b) c) Figure 52. Selection of cutting planes to divide the abdomen/pelvis of each phantom. Ovaries are illustrated in (a) as well as all of the fetal anatomy in all phantoms. Cutting planes are in orange. From left to right, a) 15, b) 25, and c) 38 week computational hybrid phantoms.

PAGE 161

161 a) b) c) Figure 5 3 Progression of mold fabrication. First the issue of interest is isolated (a) in a CAD software program. Using a series of modeling tools a computational mold is created with a vacated volume equal to the original object (b). The mold is then rendered us ing a 3D printer (c). (Photo courtesy of Matthew R. Hoerner)

PAGE 162

162 Figure 54 Density curve for ATES based on added weigh percentage of micobeads. Linear regression applied to determine the best fit equation. 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.84 0.86 0.88 0.9 0.92 0.94 0.96 0.98 Weight of microbeads added (%) Measured density (g/cc) Density = -0.238(% added) + 1.007 g/cc Regression coefficient: 0.9914

PAGE 163

163 a) b) c) Figure 55 Pregnant female anthropomorphic abdomen/pelvis phantom series. The three phantoms are a) 2/15 week hybrid b) 25 week and c) 38 week. Photo courtesy of Matthew R. Hoerner .

PAGE 164

164 Figure 56. To p ogram of the 25 week phantom (below the orange line) and the thoracic region of the adult reference female physical phantom (above the orange line). White lines at the top are from the metal cords used to bind the phantom together.

PAGE 165

165 CHAPTER 6 MEASURED CONCEPTUS DOSES ON AN ANTHROPOMORPHIC PHANTOM FROM MULTI DETECTOR COMPUTED TOMOGRAPHY STUDIES USING TUBE CURRENT MODULATION FOR ALL GESTATION AL STAGES Introduction As highlighted in Chapter 1, numerous studies have been performed to evaluate radiation absorbed dose to an unborn child from multidetector CT (MDCT). Some of these studies have employed Monte Carlo methods while others performed physical dosimetry utiliz ing anthropomorphic phantoms. In the case of Monte Carlo, none of those investigations were able to model a source term capable of evaluating absorbed fetal dose from CT studies using tube current modulation (TCM) This research is currently being done m aking computational pregnant female dosimetry certainly possible of being done in the near future.136 The physical dosimetry that has been performed, however, has never incorporated anatomical models portraying a pregnant female in the mid to late stages of pregnancy. Most have been relegated to using reference female phantoms By using adult female or male reference phantoms these studies are restricted to measuring uterine radiation doses, because none of the phanto ms have delineated locations of the developed fetus. In addition, because of individual differences in anatomy the radiation dose in the first trimester will b e different than in the second, and third. This clear lack of anatomical phantoms representing the second and third trimester prompted the investigation into radiation exposures from CT with tube current modulation techniques. E xposures from CT radiation delivered to a childbearing mother are limited to few cases. These cases include but are not limited to: pulmonary embolism (PE) acute appendicitis (AA) and trauma (T) studies These cases represent the overall majority of

PAGE 166

166 situations where the fetus receives direct exposure from radiation in CT. Other cases which use CT for diagnosis such as pancreatitis and Crohns disease are not associated with pregnancy.140 A A occurs at a rate of 0.07 0.18% during pregnancy.141, 142 Ultrasound is the preferred imaging method during the 8th to the 15th week post conception due to sensitivity of the fe tus to teratogenetic effects However, ultrasound is very user dependent and the third trimester is more difficult to perform due to the enlarged uterus and displacement of the appendix. Maternal perimeter is certainly a factor in determining the imaging approach. For a patient in an advanced stage of pregnancy, where the abdomen and pelvis have enlarged, CT provides the most information.143 Magnetic resonance imaging (MRI) can provide another option to avoid fetal exposures MRI has shown to produce comparable images for diagnosis of acute appendicitis; however, it is not the preferred method of obtaining a diagnostic image.144, 145 PE most commonly results from a blood clot in the deep veins of the legs or pelvis. It plays a significant role in maternal mort ality in the United States. One study has shown that 40% of maternal deaths are related to pulmonary embolus or cardiac disease.146 Incidence occurs between 0.05% to 0.3% of all pregnancies and is diagnosed through the use of CT angiography (CTA).147 Pregn ancy complications such as hemorrhaging of the uterus, placental abruption, placenta previa, uterine bleeding disseminated intravascular coagulopathy are leading causes of maternal mortality in the United States and require immediate medical attention.146 Clinics will often revise protocols to limit excessive radiation and preserve sufficient image quality to obtain a diagnosis. These dose reduction methods include: lowering the tube current directly or indirectly by increasing the noise factor in TCM

PAGE 167

167 studies, shortening the scan length to cover only the critical anato my, and increasing the pitch. M odern volumetric scanner s have the ability to capture entire organs and volumes of anatomy in a single rotation. These next generation scanners produce images with more coverage of anatomy in less amount of time offering new advanced procedure possibilities in noninvasive diagnostic imaging. Currently the four major vendors (Toshiba, Siemens, General Electric, and Phillips) offer sc anners with at least 64 channel (slice) configurations. Volume scanners are dynamic in thei r ability to perform studies at a variety of beam widths and slice thicknesses TCM techniques have become the standard imaging protocol for pregnant women undergoing CT scans. TCM protocols are designed to optimize dose and image quality to generate the best images possi ble without using excessive radiation. D ose and image quality are viewed as being optimized since the exposure parameter allows a tradeoff between image quality and exposure For abdominal/pelvis protocol scans the beam intensity profil e is relatively consistent over the scan due to little variation in tissue composition and effective diameter of the patient between imaging slices. For o ther exams the chest protocol for example, TCM greatly affects the x ray dose distribution within t he patient. The tube current ( mA ) ramps up around the shoulders to compensate for the amount of bone in the shoulders then decreases over the lungs (which have 33% the density of soft tissue). The second and third trimester pregnant female patient produc es a more dynamic mA(z) profile with more deviation than the first trimester mother. The circumference of the abdomen and pelvis are dependent on position along the z axis due to the protrusion of the mothers belly.

PAGE 168

168 T he absorbed radiation dose to the fetus was investigated from a series of TCM protocols using a volumetric scanner with variable image acquisition techniques to provide dosimetry information from the most modern scanners Furthermore, the absorbed doses from among all gestati onal the stages were used to a ssess the effect of TCM and acquisition settings. Materials and Methods Chapters 3 through 5 have detailed the experimental tools and measurement techniques to rapidly and accurately quantify radiation absorbed dose to the un born child in an adult reference pregnant female. In Chapters 3 and 4, two types of dosimeters were investi gated for their performance in C omputed T omography (CT) through comprehensive characterization methods. From these characterizations, these dosimet ers are on the cutting edge of CT experimental dosimetry. It is with this goal that they may be used in future studies to familiarize the radiological measurement physics community with their capabilities. In Chapter 5 a novel anthropomorphic phantom was constructed to be used with these dosimeters to provide physical anatomy necessary to evaluate radiation absorbed dose to a pregnant women spanning the period of gestational stages. The integration of these tools make it possible to collect data and eval uate the fetal exposure from CT in pregnant female studies. Anthropomorphic Phantoms The UF adult female along with the set of four abdominal/pelvis pregnant female reference phantoms were used in this study to evaluate the conceptus dose along with a se t of fiber optic dosimeters Each phantom represents an adult pregnant female with gestational ages of 2/15 week 25, and 35 weeks and based on the pregnant female hybrid computational model Each phantom represents the 50th percentile female for a

PAGE 169

169 gener al population by targeting ICRP 89 reference organ mass, height, weight, and weight gain for each stage in pregnancy.132 A more detailed description of each phantom can be found in Chapter 5. Measurement Techniques In order to evaluate the absorbed dose to the fetus, a s eries of linear fiber optic dosimeters (LF ODs ) we re used with the pregnant female physical phantom series. LFODs are valuable tools for measuring CT organ doses since they have an ability to acquire linear dose measurements which are more realistic ally represent the geometry of the target volume of interest than do conventional point dosimeters The LFOD dosimeters were calibrated from a series of measurements within a cylindrical PMMA phantom assembly Figure 6 1 illustrates the side view of the cylindrical phantom with the holes utilized for calibration and the positions of each detector relative to the primary x ray beam The detectors were calibrated by exposing the dosimeters to a helical scan over a 40cm scan length to reach dose equilibrium. In all, 2 positions were evaluated corresponding to the top and center holes The reference detector, a 0.6 cc small volume ionization chamber (Radcal, Monrovia, CA) evaluated the air kerma at each position from a helical scan with pitch equal to 1 .5 Two sets of count data were generated. One set corresponds to the primary beam conversion coefficient, while the other corresponds to the scatter radiation. Using the count rate data from the scan acquisition a scatter and primary coefficient was determined to find the number of detected counts originating from scattered radiation and primary radiation. The data was optimized iteratively to find the correction factor that relates measured signal to air kerma. For each air kerma measured, K CSig, counts wer e recorded from the plastic scintillator and CR counts from the reference fiber.

PAGE 170

170 To find the primary and scatter conversion coefficients, and respectively as well as the data was entered into a custom MatLab program that used technique to optimize the conversion coefficients by modeling the data from Equation 6 1. = , , + , , ( 6 1 ) w here the indexes 1 and 2 correspond to the hole location within the phantom. This calibration should be performed separately for each tube voltage. The signal and reference counts ( and ) are summed for each integration time above the counts originati ng from the PMT. PMT Control Program The PSD dosimetry system contains an array of five PMTs permitting both dosimeter s to be monitored simultaneously. A control program was developed in the form of a GUI (MatLab, Mathworks, Natick, MA ) to control and analyze data from both dosim e ter s simultaneously, and include input boxes for all necessary parameters to correct the light output and produce immediate dose measurements These parameters include: integration time, dead time, background correction factor, energy correction factor, and scatter and primary calibration factors. All of these parameters have been discussed in previous chapters and are highlighted in the conclusions section of Chapters 3 and 4 how they can be incorporated into a dosimetry syste m. The completed GUI is illustrated in Figure 6 2 to display all user functions and numerical inputs and outputs. Each detector requires a sampling time for all PMTs it occupies, a reference fiber correction factor for the plastic scintillator, and dead time values for all scintillators and reference fibers if desired. The program produces the measured air kerma values for each dosimeter The GUI script code is attached in Appendix A.

PAGE 171

171 CT Procedures The anthropomorphic phantoms were imaged with a Siem ens Sensation 16 scanner ( Siemens Healthcare USA, Malvern, PA ) using tube current modulation. The scanning parameters are listed for variations of routine abdomen/pelvis (appendix), CTA chest pulmonary artery (PE), and body trauma scanning protocols from UF Health radiology The anatomical landmarks defining the beginning and end of each exam are illustrated. The PE acquisition begins at the base of the adrenals and ends above the apex of the lung. The CT body trauma scan begins at the top of the kidneys or the dome of th e diaphragm and terminates at the lesser trochanter. The appendix protocol spans the length of the diaphragm to the lesser trocha n ter. Assessing Dose to the Fetus The LFOD measures the air kerma along the axis of the fetuss midline. To obtain more re alistic dosimetry measurements another LFOD w as implemented to measure the dose in the horizontal plane. One was used for the head, and the other for the chest and abdomen resulting in three air kerma measur e ments. To find a single fetal dose, the values were scaled based on the length of the each dosimeter. This is derived in Appendix B which is based on weighting each fiber by its active length. Once the total body air kerma dose was measured the radiation absorbed dose to the fetus was calculated This was done by weighting the effective mass energy absorption coefficients of soft tissue and bone component of the fetus by mass. = + ( 6 2 ) is the measured air kerma from the detector and w is the weight fraction. This is similar to the method proposed by Attix of weighting chemical compounds by

PAGE 172

172 element. M ost tissues are homogenous in nature and then typically be assigned a corresponding correction for material difference. The high magnitude of x ray energy absorption characteristics in bone leading to higher doses in CT required consideration.148 It was assumed that the bone and soft tissue are mixed together into a single entity and that the average air kerma is applicable to both. O ther studies simply just use an f factor that converts air kerma to soft tissue dose. The eff ective mass energy absorption coefficients were evaluated by using the normalized energy spectrums measured in Chapter 3 for air and polystyrene. The bone and soft tissue spectrum were equated to the polystyrene spectrums. The values are relatively indep endent of depth as demonstrated in Chapter 3. Results Absorbed Dose Measurements Using the reference values for the weight of bone and soft tissue for each gestational period, the dos e factor was calculated using Equation 62 to find the increase in radi ation absorbed dose when accounting for the amount of homogenous bone in the fetus. Equation 63 was evaluated using mass energy absorption coefficients at 50 keV for bone, soft tissue and air. The calculated values are tabulated in Table 61. Note that in the early gestational periods the fetus is comprised mostly of soft tissue, but at the later stages of pregnancy the bone begins to fully develop representing over 10% of the total fetal mass. Scans were performed using the default scan parameters for a Chest Abdomen Pelvis (CAP) exam and modifying the scan length to include the anatomy described in the above section. Table 62 shows the two protocol settings. One setting the beam energy was set to 120 kV and the other was 140 kV. For this study the tube current

PAGE 173

173 was varied by controlling the preset tubecurrent product value which directly monitored the noise level of the exam. A higher setting would result in more signal and less noise. Tables 6 3 through 6 6 show the measured fetal /embryo abs orbed doses for each of the gestational physical phantoms. The radiation absorbed dose to the 2week embryo was evaluated by taking measurements at the position of the ovaries in the 2/15week hybrid phantom. The average fetal dose for the 15th week fetus was 6.84 mGy per 100 mAs with a maximum fetal dose of 8.67 mGy per 100 mAs and a minimum of 5.45. The range of values is well below 100 m Gy, which is the established dose threshold that may induce teratogenic effects Dose measurements were normalized by the effective tubecurrent delivered to phantom from each exposure as reported by the scanner. The calculated CTDIvol was adjusted based on the effective tube current. For the 120 k Vp scans the average of the CTDIvol was 0.07026 mGy/mAs and 0.1022 mGy/mAs for 140 kVp. From the collected data, little difference in dose was seen between the CAP, T, and AA exams. The exception was for the 38 week fetus which is larger and located more superior causing it to receive more scatter radiation. The PE exam was completely out of the scan field and resulted in small amounts of accumulated dose. Dose Estimation Models Figure 63 plotted the relationship between CTDIvol and fetal dose. Curves are linear but display different fitting functions. Each curve contains data from both energies and mA settings. For h igh scanner outputs the values begin to deviate from one another and a single model becomes difficult. Each data set was fit to a first degree polynomial functions. The fit for the 25week phantom is y = 1.0554x 0.71948, and the fit for the 15week phantom is y = 0.8465x + 0.11707. The linear nothreshold

PAGE 174

174 fit is 1.0019 mGy/mGy (R2 = 0.9921) and 0.8561 mGy/mGy (R2 = 0.9820) for the 25 and 15 week phantoms respectively. This is more of an ideal fit since it has physical meaning since if the tube current was set to 0 then the CTDI would be 0 and so would the fetal dose. Another interesting relationship that is often not documented is the relationship between dose and scan length. Since data is acquired by the detection system in real time and is recorded a count vs. time profile can be generated. Since the x ray tube moves along the z axis counts vs. position plot can be formulated to show accumulated dose vs. scan length. Figure 64 show s the total radiation absorbed dose as a function of scan length (moving inferior to superior and vi ce versa) for the 15 and 25 week phantoms at 120 kV. The data shows that the developed fetus is more sensitive to changes in scan length which can be attri buted to more scatter being generated. Before 200 mm on a CAP exam (head first) no dose is received by the fetus. This Figure may help clinics set protocols for PE exams which mainly focus on the chest through the liver (pulmonary artery). Discussion This study presented unique experimental methods of evaluating fetal radiation absorbed dose to multiple gestations from tube current modulated scanner techniques also show ing that caution should be exercised when using the CTDI size specific correction f actors found in literature. The conversion coefficient between absorbed dose and CTDIvol for the 25week phantom was higher in magnitude than the 15week phantom. This contradicts the AAPM report that uses sizespecific correction factors to correlate CT DIvol and absorbed dose which found that a larger perimeter resulted in a lower dose due to attenuation. However, a significant factor in the increase in absorbed

PAGE 175

175 dose is the increase of the weight fraction of bone within the fetus. When using TCM t he e ffects of patient size are minimized because the x ray fluence attenuated by the patient will be compensated by the x ray tube to maintain a consistent signal level hitting detector array. Also, CTDIvol reported for this scanner is based on the average mA not the mA irradiating the fetus which is probably the maximum mA given that the mothers waist is at the largest circumference at the fetuss z axis position. Another interesting discovery was that at older gestational ages are more sensitive to changes in the mA setting. The 15week phantom absorbed radiation dose is approximately 0.047 mGy per mAs at 120 kV, compared to the 0.082 mGy per mAs of the 25week phantom at 140 kV from a CAP exam This is mostly likely due to the scatter radiation contribut ion near the fetus. The advantage of using LFODs and these anthropomorphic phantoms is evaluating organs located on the periphery on a big cross section. The radiation dose is very dependent on the angle of the tube as it transverses over the organ. A t anterior angles lower doses will be obtained and posterior angles will produce higher doses. Point dosimeters placed in a cross sectional phantom will only monitor the dose for that single plane. If the organ is of a length that the tube will make it back around those doses will be not captured and the dose will be inaccurate. If multiple rotations are made around the organ than the dose readings will average out and approach the actual organ dose. These circumstances are dependent on the exam pitch, organ length, and slice thickness. Additionally, if the dosimeters are placed at equal distances apart in the z axis then there is a probability that the beam will intersect each dosimeter in the same location and the true organ dose value will never be c aptured. These

PAGE 176

176 issues would never be significant using a LFOD placed along the organs length in the z axis because measurements are obtained along the entire length. T he calibration methods being implemented by the literature discussing CT dosimetry ar e subject to scrutiny Free in air calibrations will not account for the scatter radiation exposed to the detector. All three studies that performed physical measurements calibrated their TLDs freein air. Scatter radiation is a different energy than pr imary x ray radiation and will have a different correction factor. Most of the fetal dose measurements were obtained in the central locations of the phantom, where scatter is a significant contribution to the total absorbed dose. One of the limitations to this study was obtaining data from a phantom simulating a single body habitus with a conceptus oriented in a fixed position. Radiation doses vary depending on the size, fat/muscle composition of the phantom, and the position of the fetus within the mot her. Preliminary sensitivity studies using Monte Carlo have placed the uncertainty in measuring doses from varying soft tissue density to be anywhere between 0.0 4 to 0.0 6 mGy per 0.01 g/cc. However, this value was calculated from simulations performed on a n adult reference male. Therefore a soft tissue density of 1.02 g/cc would result in an additional radiation absorbed dose of 0.12 mGy (1% difference) It would be expected that the larger pelvis diameter would have a higher magnitude of an effect. Exact doses will depend on the imaging protocols established by the institution. Imaging techniques to lower the absorbed dose that are applicable to MDCT include but are not limited to: lower ing the tube current, minimizing the scan length to reduce scatter contributions, and increasing the pitch. Increasing the pitch and lowering

PAGE 177

177 the tube current may result in poorer image quality. Modifying the scan length would have a greater effect on regions in the center of the patient which receive more dose con tribution from scatter radiation. Changing the pitch may or may not have an effect on regions located on the periphery because the primary beam may still directly pass through that region. This is especi ally true for smaller organs where the primary bea m transverses the plane of the region in a single pass, which are more likely to occur at wider slice thicknesses. A potential technique to lower the fetal dose in MDCT that is applicable to volumetric scanning would be to raise the pitch and optimize th e starting angle to control the tube position. This would create a scenario where the direct x ray beam passes th rough the fetus mostly from the posterior surface of the mother s skin and less from the anterior surface. Although currently this cannot be done since the starting angle is uncontrollable, it poses as interesting and potential technique to monitor exposures in CT. Conclusion Radiation doses to the embryo and fetus from the appendix and body trauma TCM scans are shown to contribute a relatively equal amount of dose that falls below the level of concern. It was expected that the PE protocol would yield lower radiation doses than the other protocols because the fetus/embryo is only exposed to scatter radiation The fetal dose from a PE exam i s most dependent on the scan length in the inferior direction of the patient. The basis of these dose estimates show that the doses are highly dependent on the exam protocols especially the tube potential. Clinics that use 140 kV tube voltage and high tube current settings will irradiate the fetus at a much higher level than clinics using low dose protocols. Some of the results of this study are

PAGE 178

178 applicable to large patient dosimetry and could result in optimizing exam protocols to minimize excessive dose. Finally, this study can potentially serve as a basis to develop a clinically practiced fetal dose calculation methodology from its empirical finds as well as its potential to validate Monte Carlo TCM simulations for various CT scanners.

PAGE 179

179 T able 6 1. Dose c oefficients for converting fetal soft tissue doses to total body doses. Gestation Bone mass (g) Total fetal mass (g) / ratio of bone to soft tissue Total body dose coefficient 15 week < 1 160 1.105 1.00 25 week 56.85 990 1.567 1.09 35 week 293.06 2700 2.486 1.27 Table 62. Imaging parameters for 16MDCT scanner. Parameter Protocol A Protocol B Peak kilovoltage (kV) 120 140 Detector configuration 16 x 1.5 mm 16 x 1.5 mm Pitch 0.75:1 0.75:1 Tube rotation (sec) 0.5 0.5 Normalized CTDIvol (mGy/mAs) 0.0703 0. 102 Avg. s can length: CAP (mm) 66 1 661 Avg. s can length: AA (mm) 36 0 36 0 Avg. s can length: T (mm) 29 0 29 0 Avg.scan length: PE (mm) 33 0 33 0

PAGE 180

180 Table 6 3 Radiation absorbed embryo doses from a variety of pelvic exams Radiation absorbed fetal dose (mGy) Avg. T/AA/CAP absorbed dose [mGy] TCM setting (mAs CAP AA T PE Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B 280 10.89 13.68 10.78 13.47 10.46 13.07 0.20 0.23 10.71 13.41 240 9.16 11.97 9.07 11.79 8.90 11.52 0.09 0.18 9.04 11.76 200 8.21 10.66 8.11 10.56 7.94 10.31 0.13 0.12 8.09 10.51 160 6.52 8.74 6.29 8.67 6.26 8.59 0.09 0.08 6.36 8.67 Table 64 Radiation absorbed 15week fetal doses from a variety of pelvic exams Radiation absorbed fetal dose (mGy) Avg. T/AA/CAP absorbed dose [mGy] TCM setting (mAs CAP AA T PE Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B 280 12.30 15.47 12.18 15.23 11.83 14.83 0.21 0.25 12.10 15.18 240 9.61 13.21 9.52 13.02 9.35 12.74 0.08 0.20 9.49 12.99 200 8.20 11.24 8.10 11.12 7.95 10.89 0.13 0.12 8.08 11.08 160 6.50 8.83 6.29 8.75 6.26 8.67 0.09 0.09 6.35 8.75

PAGE 181

181 Table 65 Radiation absorbed 25week fetal doses from a variety of pelvic exams. Radiation absorbed fetal dose (mGy) Avg. T/AA/CAP TCM setting (mAs) CAP AA T PE absorbed dose [mGy] Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B 280 16.98 20.85 16.80 20.44 16.37 19.81 0.19 0.29 16.72 20.37 240 14.45 15.78 14.29 16.41 13.91 15.83 0.17 0.25 14.22 16.01 200 11.76 12.48 11.61 13.11 11.27 12.70 0.19 0.19 11.55 12.76 160 9.30 9. 94 9.19 9.80 8.87 9.49 0.14 0.14 9.12 9.74 Table 66. Radiation absorbed 38week fetal doses from a variety of pelvic exams. Radiation absorbed fetal dose (mGy) Avg. T/AA/CAP TCM setting (mAs) CAP AA T PE absorbed dose [mGy] Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B Prot. A Prot. B 280 19.03 27.88 18.42 26.99 16.97 24.60 0.61 0.88 18.14 26.49 240 16.09 22.98 15.56 22.25 14.09 20.70 0.54 0.72 15.25 21.98 200 13.51 19.18 13.06 18.61 12.00 17.12 0.46 0.58 12.86 18.30 160 10.80 15.11 10.48 14.64 9.64 13.58 0.32 0.47 10.31 14.44

PAGE 182

182 Figure 61. Phantom assembly used to calibrate the dosimeter s. Two CTDI 32 cm body phantoms and three soft tissueequivalent phantoms provided over 40 cm of build up to reach dose equilibrium along the central axis. (Photo courtesy of Matthew R. Hoerner )

PAGE 183

183 Figure 62. MATLAB GUI dosimetry program with user functions and output data.

PAGE 184

184 Figure 63. Relationship between fetal dose and CTDIvol at various gestational ages for a CAP examination 6 8 10 12 14 16 18 20 5 10 15 20 25 30 CTDIvol (mGy)Radiation absorbed fetal dose (mGy) 15 week 25 week 38 week 15 week fit 25 week fit 38 week fit

PAGE 185

185 Figure 64. Accumulated fetal dose as function of scan position. Data taken from the CAP scans (head first) using protocol A. 0 100 200 300 400 500 600 700 0 2 4 6 8 10 12 14 Accumulated dose (mGy)Scan length (mm) 15 week 25 week 38 week

PAGE 186

186 CHAPTER 7 CONCLUSIONS AND FUTURE WORK Plastic Fiber Optic Detectors General Notes The effectiveness of PSDs as a dosimetry tool is dependent on many parameters such as the particle energy, exposure rate, photon detector, and exposure of the optical fiber From a diagnostic energy perspective the effect of quenching is very apparent and it seems it may be more productive to explore other types of scintillat ion materials This is especially pertinent to the development of a dual scintillating dosimeter which would be more highly effective using scintillators that are not prone to quenching in order to better resolve the beam energy spectrum and make appropriate corrections. This will be discussed in further detail in the future work section. When using these detectors for megavoltage x ray and particle beams a noticeable field size and dose rate is apparent. However it was demons trated to be correctable, though not ideal for performing accurate dosimetry. When looking at other studies it appears that these issues stem from the photon detectors time response from pulse counting compared to signal integration. This would suggest moving towards a more conventional photon detector, such as a CCD camera that integrates the signal This work is currently being don e at University of Florida for ConeB eam CT (CBCT) dosimetry applications. T his dissertation has clearly demonstrated the correction methods for count rate effects stemming from dead time for organic and inorganic scintillators using a photomultiplier tube array. Using the correction method detailed in this document provides a un iversal technique for characterizing PSDs for any type of radiation detection and measurement application. The direct relationship between dead time and

PAGE 187

187 Cerenkov radiation is not definite but seems very likely. Our analysis demonstrates that the irradiat ion of an optical fiber and possibly scintillator create light attenuation/absorption issues that may be more or less magnified depending on the type of photon detector used. When transitioning from the PMT to the CCD, the count rate effects and stem effects from high energy radiation should be characterized and applied directly to the dead time model in this manuscript. It should be noted however that the nonparalzable model is an ideal and not applicable to all types of detectors. Linear fiber opti c dosimeters offer the most utility for performing organ dose measurements in the radiological energy range owing to their spatially integrated dose measurement The application of one dimensional dosimetry is relatively new and presents more opportunities to expand the capabilities of these devices into more medical modalities. In addition, the real time monitoring of counts from helical scanning has shown promise in reducing the energy dependence of a detector by measuring the scatter and primary x ray beam contribution. The limited cases presented in this study show that identifying the scatter and primary radiation contribution can self correct the detector for energy and depth dependence to perform accurate dosimetry. Energy Dependence of a Plast ic Scintillator The results obtained in this dissertation have produced a general notion to the response of a plastic scinti llator to irradiation from radiological x rays. Figure 8 1 illustrates the mass energy absorption coefficients of polystyrene to air for energies ranging from 1 to 150 keV. A linear fit (R = 0.9986) was applied to the data between 35 and 75 keV to demonstrate the linearity of the energy dependence between those energies. For diagnostic x ray beams characterized by an effective energy between 35 keV and 75

PAGE 188

188 keV, any change to that spectrum would increase or decrease the energy dependence of the detector proportional to a constant. The linearity of the energy dependence was characterized in this manuscript as well as Hyer et al. The energy dependence of a plastic scintillator in CT can be described empirically as a function of both beam energy depth. Equation 81 formulates an expression for the correction factor of a plastic scintillator to appropriately correct for the energy dependence. = ( ( ) + ) + ( 7 1 ) Where and are the fraction of counts that are a result of interactions from the primary radiation and scatter radiation with the scintillator. E is defined as the effective energy of the x ray beam freein air. The primary beam coefficients ( and ) are model the energy dependence coefficient from changes to tube potential and depth. Equation 81 could serve as a model for correcting the response of the detector to the response of an ioni zation chamber. These investigations have shown that plastic scintillators used in radiological dosimetry demonstrate a quenching effect. Further evidence of quenching is supported by the slope of Figure 81 compared to the slope of the plastic scintill ating dosimeter. From Figure 46 the slope of that line when converting from kVp to effective energy of each spectrum is 0.021 per keV. This value is double the 0.0112 per keV value from applying large cavity theory. This may explain why Lessard et al. was able to use large cavity theory to correct for the response of the detector without accounting for quenching. The quenching parameter can be hidden within the scintillating efficiency or any other conversion parameter in the 60140 kVp energy range. The work of Frenlin et al. showed that quenching is linear with respect to energy in this energy range. When

PAGE 189

189 accounting for quenching and mass energy absorption, the sensitive element demonstrates a linear energy dependence from direct radiation in CT. W hen scatter is present this linearity breaks down as evidence by the results in Chapter 3, and the right term in Equation 81 dominates the energy dependence. The magnitude of the quenching effect observed by the BCF 12 plastic scintillators used in our d etectors appears to be significantly higher than those of other studies. The cause of this is largely unknown. Anthropomorphic Phantoms There is a great benefit of rendering phantom molds using 3D printers compared to the traditional method using a 2D engraver. However, using a 3D printer brings a higher cost associated with building each mold. In addition constructing molds is complicated by the limiting build volume of a 3D printer (10 x 8 x 8) compared to the surface area of a 2D engraver (24 x 16). Building anthropomorphic phantoms means the full cross sections of adults must be composed of multiple components, which adds to the build cost and construction time. Considerations must include the ease of extracting the soft tissue material from the mold and the marginal cost associated with each mold. For this project 275.55 in3 of material was used at an estimated cost of $826.65 compared to the approximate $100 of foam board that would be required for this project. However, more tissue mater ial would have been needed because each phantom would have been built separately instead of into components which helps alleviate the overall cost of the project. Physical Dosimetry One of the findings that these investigations have founded is that scatt er radiation plays a prominent role in the radiation absorbed dose received from CT

PAGE 190

190 exposures This scatter radiation is a completely different energy and induces a completely different detector response than primary radiation. When using a detector with a sensitive element that is greater than that of air or soft tissue its correction factor will be less than 1. At low energies the correction factor will be the smallest and at higher energies it will start to approach 1. If the detector is calibrated f ree in air it will overestimate the correction factor because the scatter contribution, which has a lower coefficient, is ignored. In this case CT doses are overestimated. When using a detector with a smaller effective Z, the reverse case is true and radiation absorbed doses will be underestimated. The calibration method presented in this study is ideal for CT dosimetry because it will properly simulate the scatter and primary radiation experienced by each detector. Using a CTDI phantom and a single rot ation non table translation will be a closer approximation, as it produces an estimated 2/3 primary and 1/3 scatter. However, for large patients the actual value for a CT scan approaches equal contributions of both. In any case, it is a much better estim ation than freein air calibration methods. A review of the detectors described in Chapter 3 used the dosimetry studies presented are LiF TLDs ( TLD 100, Harshaw, Solon, OH) and MOSFET ( 1002RD, Best Medical Canada, Ottawa, Ontario) dosimeters. The MOSFET sensitive element is a dual transistor with one element made of Silicon and the other SiO2. Figure 72 plots the energy response of these two detectors along with the PSD relative to air to demonstrate their energy response correction factors at high and low keV energies The TLD shares a similar energy response to the MOSFET dosimeter relative to air. Therefore, when both are calibrated freein air their response will be over responsive

PAGE 191

191 and overestimate the radiation absorbed dose. Although, the TLD seems to much less affected since the correction factors only varies less across the energy range of interest. Future Work Future Work Project A This project focuses on constructing a PMT photon counter device that reduces the dead time, and more importantly, analyzes the counting losses that it the system is experiencing. The PMT overall is a fast responsive device that converts light into an electrical current. The conversion of charge into pulses is what ultimately limits the response time. The phot on counter of the detector array studied in this work consists of an amplifier, comparator, pulse shaper, and counter. All of these devices are relatively inexpensive and with the circuit boards our department owns would allow our lab group to construct o ur own device; permitted we have an accessible PMT. The first part of this project would be to analyze the output of the PMT from different radiation sources to simulate the response of the PMT. These responses can be simulated from a function generator and connected to a circuit and viewed with an oscilloscope to decipher any issues with photon counting. Dr. James Schumacher from the University of Florida, is a kn owledgeable source and his lab has almost of all of the equipment that would be needed. The research is based entirely on constr u cting a circuit that optimizes the signal processing for photon detection in real time. Considerations should include: photon counting vs signal integration, and time response of the electrical components. Future Work Project B This project is focused on rebuilding the dual scintillating detector using two inorganic scintillator phosphors, of which neither of demonstrate quenching effects.

PAGE 192

192 These scintillators could be extracted from an x ray film cassette or fr om a vendor. Ideally the use of low Z scintillators would be used, which have K edge below the common energies encountered in CT. However high k edges have proven useful as well since most of the absorbed energy coincides with the effective energy of the x ray and sensitivity to scatter which make their response less variable. The methods for characterizing such a detector are laid out in Chapter 3. In addition to this project, it would of interest to compare each detectors freein air and in a scatter ing medium to quantify the over response in reported doses by previous studies. It would also be useful to evaluate the energy dependence across a range of tube potentials (20 thru 200 kVp) to demonstrate the effects of quenching outside the 60 thru 140 k Vp range. Future Work Project C Finally, there is now a physical validation methodology in place to verify Monte Carlo simulations on the computational pregnant female series utilizing TCM. At this moment Dr. Bolchs research group has already created a TCM source term for a Toshiba Aquilion ONE scanner. One of the limits of the source model however, is that it can only perform retrospective dosimetry since the actual TCM algorithm is proprietary. Therefore images from the real CT scans are r equired to report the average/effective mAs per slice, which can be used to reconstruct the source profile and be used for the simulated dosimetry. Measurements following the methodology in this dissertation would be repeated on the scanner and given to D r. Bolchs research group to perform the Monte Carlo calculations. This would be a great project since both the data necessary for dosimetry, scanner rep o rted CTDIvol values and DLP values, and physical measurements can be acquired by our research group that would vital

PAGE 193

193 potentially leading to a general dosimetry model and validation of some of the results in this manuscript.

PAGE 194

194 Figure 7 1. Plot of / ratio between polystyrene and air in the low energy range, taken from the National Institute of Standards and Technology (Ref 108). Linear fit was applied to the data points between 35 and 75 keV. 0 50 100 150 0 0.25 0.5 0.75 1 1.25 1.5 1.75 Photon energy (keV) Mass energy-absorp ratio (poly to air) y = 0.0115x

PAGE 195

195 Figure 72. Plot of / ratio between va rious dosimeters and air in the low energy range, taken from the National Institute of Standards and Technology (Ref 108). LiF is the sensitive element of the TLD and silicon dioxide is that of the MOSFET dosimeter. 0 50 100 150 0 0.5 1 1.5 2 2.5 3 Photon energy (keV)Dosimeter mass energy-absorp. ratio (rel. to air) LiF SiO2 Poly

PAGE 196

196 APPENDIX A MATLAB CODE FOR GRAPHICAL INTERFACE function varargout = CT_PMT(varargin) % CT_PMT Mfile for CT_PMT.fig % CT_PMT, by itself, creates a new CT_PMT or raises the existing % singleton*. % % H = CT_PMT returns the handle to a new CT_PMT or the handle to % the existing singleton*. % % CT_PMT('CALLBACK',hObject,eventData,handles,...) calls the local % function named CALLBACK in CT_PMT.M with the given input arguments. % % CT_PMT('Property','Value',...) creates a new CT_PMT or raises the % existing singleton*. Starting from the left, property value pairs are % applied to the GUI before CT_PMT_OpeningFcn gets called. An % unrecognized property name or invalid value makes property application % stop. All inputs are passed to CT_PMT_OpeningFcn via varargin. % % *See GUI Options on GUIDE's Tools menu. Choose "GUI allows only one % instance to run (singleton)". % % See also: GUIDE, GUIDATA, GUIHANDLES % Edit the above text to modify the response to help CT_PMT % Last Modified by GUIDE v2.5 22Jan2014 00:09:29 % Begin initialization code DO NOT EDIT gui_Singleton = 1; gui_State = struct('gui_Name', mfilename, ... 'gui_Singleton', gui_Singleton, ... 'gui_OpeningFcn', @CT_PMT_OpeningFcn, ... 'gui_OutputFcn', @CT_PMT_OutputFcn, ... 'gui_LayoutFcn', [] ... 'gui_Callback', []); if nargin && ischar(varargin{1}) gui_State.gui_Callback = str2func(varargin{1}); end if nargout [varargout{1:nargout}] = gui_mainfcn(gui_State, varargin149); else gui_mainfcn(gui_State, varargin149); end % End initialization code DO NOT EDIT % --Executes just before CT_PMT is made visible. % DEFAULT FUNCTION function CT_PMT_OpeningFcn(hObject, eventdata, handles, varargin) % This function has no output args, see OutputFcn.

PAGE 197

197 % hObject handle to figure % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % varargin command line arguments to CT_PMT (see VARARGIN) % Choose default command line output for CT_PMT handles.output = hObject; % Update handles structure handles.data.count = 1; set(hObject,'toolbar', 'figure'); %sets defalt toolbar onto GUI (top of figure) guidata(hObject, handles); % UIWAIT makes CT_PMT wait for user response (see UIRESUME) % uiwait(handles.figure1); % --Outputs from this function are returned to the command line. % DEFAULT FUNCTION function varargout = CT_PMT_OutputFcn(hObject, eventdata, handles) % varargout cell array for returning output args (see VARARGOUT); % hObject handle to figure % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Get default command line output from handles structure varargout{1} = handles.output; %%%% Integration Time Input EDIT Text Boxes f unction integration_time_input_PMT1_Callback(hObject, eventdata, handles) % hObject handle to integration_time_input_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of integration_time_input_PMT1 as text % str2double(get(hObject,'String')) returns contents of integration_time_input_PMT1 as a double handles.data.int_time_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. % edit text box for PMT 1 to input integration time function integration_time_input_PMT1_CreateFcn(hObject, eventdata, handles) % hObject handle to integration_time_input_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER.

PAGE 198

198 if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.int_time_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on integration_time_input_PMT1 and none of its controls. function integration_time_input_PMT1_KeyPressFcn(hObject, eventdata, handles) % hObject handle to integration_time_input_PMT1 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function integration_time_input_PMT2_Callback(hObject, eventdata, handles) % hObject handle to integration_time_input_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of integration_time_input_PMT2 as text % str2double(get(hObject,'String')) returns contents of integration_time_input_PMT2 as a double handles.data.int_time_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function integration_time_input_PMT2_CreateFcn(hObject, eventdata, handles) % hObject handle to integration_time_input_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.int_time_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on integration_time_input_PMT2 and none of its controls. function integration_time_input_PMT2_KeyPressFcn(hObject, eventdata, handles) % hObject handle to integration_time_input_PMT2 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed

PAGE 199

199 % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) %%%%% Dead Time Edit Text Boxes function dead_time_input_PMT1_Callback(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of dead_time_input_PMT1 as text % str2double(get(hObject,'String')) returns contents of dead_time_input_PMT1 as a double handles.data.dead_time_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function dead_time_input_PMT1_CreateFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.dead_time_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on dead_time_input_PMT1 and none of its controls. function dead_time_input_PMT1_KeyPressFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT1 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function dead_time_input_PMT2_Callback(hObject, eventdata, handles) % hObject handle to codead_timeunting_threshold_input_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of dead_time_input_PMT2 as text % str2double(get(hObject,'String')) returns contents of dead_time_input_PMT2 as a double handles.data.dead_time_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles);

PAGE 200

200 % --Executes during object creation, after setting all properties. function dead_time_input_PMT2_CreateFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.dead_time_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on dead_time_input_PMT2 and none of its controls. function dead_time_input_PMT2_KeyPressFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT2 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function dead_time_input_PMT3_Callback(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT3 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of dead_time_input_PMT3 as text % str2double(get(hObject,'String')) returns contents of dead_time_input_PMT3 as a double handles.data.dead_time_PMT3=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function dead_time_input_PMT3_CreateFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT3 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.dead_time_PMT3=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on dead_time_input_PMT3 and none of its controls. function dead_time_input_PMT3_KeyPressFcn(hObject, eventdata, handles)

PAGE 201

201 % hObject handle to dead_time_input_PMT3 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function dead_time_input_PMT4_Callback(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT4 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of dead_time_input_PMT4 as text % str2double(get(hObject,'String')) returns contents of dead_time_input_PMT4 as a double handles.data.dead_time_PMT4=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function dead_time_input_PMT4_CreateFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT4 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.dead_time_PMT4=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on dead_time_input_PMT4 and none of its controls. function dead_time_input_PMT4_KeyPressFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT4 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function dead_time_input_PMT5_Callback(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT5 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of dead_time_input_PMT5 as text % str2double(get(hObject,'String')) returns contents of dead_time_input_PMT5 as a double handles.data.dead_time_PMT5=str2double(get(hObject,'String'));

PAGE 202

202 guidata(hObject, handles); % --Executes during object creation, after setting all properties. function dead_time_input_PMT5_CreateFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT5 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.dead_time_PMT5=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on dead_time_input_PMT5 and none of its controls. function dead_time_input_PMT5_KeyPressFcn(hObject, eventdata, handles) % hObject handle to dead_time_input_PMT5 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) %%%%%%%%%% REF CALIBRATION FACTORS %%%%%%%%%%%%%%%%%% function ref_cal_fact_PMT1_Callback(hObject, eventdata, handles) % hObject handle to ref_cal_fact_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of ref_cal_fact_PMT1 as text % str2double(get(hObject,'String')) returns contents of ref_cal_fact_PMT1 as a double handles.data.ref_cal_fact_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function ref_cal_fact_PMT1_CreateFcn(hObject, eventdata, handles) % hObject handle to ref_cal_fact_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.ref_cal_fact_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles);

PAGE 203

203 % --Executes on key press with focus on ref_cal_fact_PMT1 and none of its controls. function ref_cal_fact_PMT1_KeyPressFcn(hObject, eventdata, handles) % hObject handle to ref_cal_fact_PMT1 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function ref_cal_fact_PMT2_Callback(hObject, eventdata, handles) % hObject handle to ref_cal_fact_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of ref_cal_fact_PMT2 as text % str2double(get(hObject,'String')) returns contents of ref_cal_fact_PMT2 as a double handles.data.ref_cal_fact_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function ref_cal_fact_PMT2_CreateFcn(hObject, eventdata, handles) % hObject handle to ref_cal_fact_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.ref_cal_fact_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on ref_cal_fact_PMT2 and none of its controls. function ref_cal_fact_PMT2_KeyPressFcn(hObject, eventdata, handles) % hObject handle to ref_cal_fact_PMT2 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) %%%%%%%%%% PRIMARY BEAM CALIBRATION FACTORS %%%%%%%%%%%%%%%%%% function prim_cal_fact_PMT1_Callback(hObject, eventdata, handles) % hObject handle to prim_cal_fact_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA)

PAGE 204

204 % Hints: get(hObject,'String') returns contents of prim_cal_fact_PMT1 as text % str2double(get(hObject,'String')) returns contents of prim_cal_fact_PMT1 as a double % --Executes during object creation, after setting all properties. handles.data.prim_cal_fact_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); function prim_cal_fact_PMT1_CreateFcn(hObject, eventdata, handles) % hObject handle to prim_cal_fact_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.prim_cal_fact_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on prim_cal_fact_PMT1 and none of its controls. function prim_cal_fact_PMT1_KeyPressFcn(hObject, eventdata, handles) % hObject handle to prim_cal_fact_PMT1 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function prim_cal_fact_PMT2_Callback(hObject, eventdata, handles) % hObject handle to prim_cal_fact_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of prim_cal_fact_PMT2 as text % str2double(get(hObject,'String')) returns contents of prim_cal_fact_PMT2 as a double handles.data.prim_cal_fact_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function prim_cal_fact_PMT2_CreateFcn(hObject, eventdata, handles) % hObject handle to prim_cal_fact_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor'))

PAGE 205

205 set(hObject,'BackgroundColor', 'white'); end handles.data.prim_cal_fact_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on prim_cal_fact_PMT2 and none of its controls. function prim_cal_fact_PMT2_KeyPressFcn(hObject, eventdata, handles) % hObject handle to prim_cal_fact_PMT2 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) %%%%%%%%%%%%%%%%%%SECONDARY CALIBRATION FACTORS %%%%%%%%%%%%%%%%% function sec_cal_fact_PMT1_Callback(hObject, eventdata, handles) % hObject handle to sec_cal_fact_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of sec_cal_fact_PMT1 as text % str2double(get(hObject,'String')) returns contents of sec_cal_fact_PMT1 as a double handles.data.sec_cal_fact_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function sec_cal_fact_PMT1_CreateFcn(hObject, eventdata, handles) % hObject handle to sec_cal_fact_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.sec_cal_fact_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on sec_cal_fact_PMT1 and none of its controls. function sec_cal_fact_PMT1_KeyPressFcn(hObject, eventdata, handles) % hObject handle to sec_cal_fact_PMT1 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA)

PAGE 206

206 function sec_cal_fact_PMT2_Callback(hObject, eventdata, handles) % hObject handle to sec_cal_fact_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of sec_cal_fact_PMT2 as text % str2double(get(hObject,'String')) returns contents of sec_cal_fact_PMT2 as a double handles.data.sec_cal_fact_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function sec_cal_fact_PMT2_CreateFcn(hObject, eventdata, handles) % hObject handle to sec_cal_fact_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.sec_cal_fact_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on sec_cal_fact_PMT2 and none of its c ontrols. function sec_cal_fact_PMT2_KeyPressFcn(hObject, eventdata, handles) % hObject handle to sec_cal_fact_PMT2 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) %%%%%%%%%%%%%%%%%%%%%%%%% COUNT RATE THRESHOLD %%%%%%%%%%%%%%%%%%%%%%%%%% function cr_thresh_PMT1_Callback(hObject, eventdata, handles) % hObject handle to cr_thresh_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of cr_thresh_PMT1 as text % str2double(get(hObject,'String')) returns contents of cr_thresh_PMT1 as a double handles.data.cr_thresh_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function cr_thresh_PMT1_CreateFcn(hObject, eventdata, handles) % hObject handle to cr_thresh_PMT1 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB

PAGE 207

207 % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.cr_thresh_PMT1=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on cr_thresh_PMT1 and none of its controls. function cr_thresh_PMT1_KeyPressFcn(hObject, eventdata, handles) % hObject handle to cr_thresh_PMT1 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function crt_thresh_PMT2_Callback(hObject, eventdata, handles) % hObject handle to crt_thresh_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of crt_thresh_PMT2 as text % str2double(get(hObject,'String')) returns contents of crt_thresh_PMT2 as a double handles.data.crt_thresh_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function crt_thresh_PMT2_CreateFcn(hObject, eventdata, handles) % hObject handle to crt_thresh_PMT2 (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.crt_thresh_PMT2=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on crt_thresh_PMT2 and none of its controls. function crt_thresh_PMT2_KeyPressFcn(hObject, eventdata, handles) % hObject handle to crt_thresh_PMT2 (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed

PAGE 208

208 % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) %%%%%%%%%%%%%%%%% DUAL SCINT CONTROLS %%%%%%%%%%%%%%%%%%%%%%% function center_ratio_Callback(hObject, eventdata, handles) % hObject handle to center_ratio (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of center_ratio as text % str2double(get(hObject,'String')) returns contents of center_ratio as a double handles.data.center_ratio=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function center_ratio_CreateFcn(hObject, eventdata, handles) % hObject handle to center_ratio (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.center_ratio=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on center_ratio and none of its controls. function center_ratio_KeyPressFcn(hObject, eventdata, handles) % hObject handle to center_ratio (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) function ratio_cal_fact_Callback(hObject, eventdata, handles) % hObject handle to ratio_cal_fact (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of ratio_cal_fact as text % str2double(get(hObject,'String')) returns contents of ratio_cal_fact as a double handles.data.ratio_cal_fact=str2double(get(hObject,'String'));

PAGE 209

209 guidata(hObject, handles); % --Executes during object creation, after setting all properties. function ratio_cal_fact_CreateFcn(hObject, eventdata, handles) % hObject handle to ratio_cal_fact (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.ratio_cal_fact=str2double(get(hObject,'String')); guidata(hObject, handles); % --Executes on key press with focus on ratio_cal_fact and none of its controls. function ratio_cal_fact_KeyPressFcn(hObject, eventdata, handles) % hObject handle to ratio_cal_fact (see GCBO) % eventdata structure with the following fields (see UICONTROL) % Key: name of the key that was pressed, in lower case % Character: character interpretation of the key(s) that was pressed % Modifier: name(s) of the modifier key(s) (i.e., control, shift) pressed % handles structure with handles and user data (see GUIDATA) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % PMT/USER Control Buttons % --Executes on button press in Start. function Start_Callback(hObject, eventdata, handles) % hObject handle to Start (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Set Start to value of 1 so as long as activated data will be aquired set(handles.Start,'UserData',1); %integration time in binary format to send to each available PMT inttime_1=uint8([128,0,0,0,handles.data.int_time_PMT1/10]); inttime_2=uint8([128,0,0,0,handles.data.int_time_PMT2/10]); if handles.data.count == 1 % open up comnication with PMT's 1 thru 5 handles.data.count = 0; measurenum = uint8([129,0,0,0,1]); % measurement number to send to the PMT (usually 1) handles.data.startcmd = uint8([132,0,0,0,1]); % command to begin counting sent to PMT handles.data.stopcmd = uint8([132,0,0,0,0]); % command to stop counting sent to PMT

PAGE 210

210 %PMT 1 % open serial port to PMT 1 handles.data.serialobj1=serial('COM101', 'BaudRate', 9600, 'Parity', 'none', 'InputBufferSize', 30000); fopen(handles.data.serialobj1); fwrite(handles.data.serialobj1,handles.data.stopcmd); fwrite(handles.data.serialobj1,measurenum); fwrite(handles.data.serialobj1,inttime_1); fwrite(handles.data.serialobj1,handles.data.startcmd); %PMT 2 % open serial port to PMT 2 handles.data.serialobj2=serial('COM102', 'BaudRate', 9600, 'Parity', 'none', 'InputBufferSize', 30000); fopen(handles.data.serialobj2); fwrite(handles.data.serialobj2,handles.data.stopcmd); fwrite(handles.data.serialobj2,measurenum); fwrite(handles.data.serialobj2,inttime_1); fwrite(handles.data.serialobj2,handles.data.startcmd); %PMT 3 % open serial port to PMT 3 handles.data.serialobj3=serial('COM103', 'BaudRate', 9600, 'Parity', 'none', 'InputBufferSize', 30000); fopen(handles.data.serialobj3); fwrite(handles.data.serialobj3,handles.data.stopcmd); fwrite(handles.data.serialobj3,measurenum); fwrite(handles.data.serialobj3,inttime_1); fwrite(handles.data.serialobj3,handles.data.startcmd); %PMT 4 % open serial port to PMT 4 handles.data.serialobj4=serial('COM104', 'BaudRate', 9600, 'Parity', 'none', 'InputBufferSize', 30000); fopen(handles.data.serialobj4); fwrite(handles.data.serialobj4,handles.data.stopcmd); fwrite(handles.data.serialobj4,measurenum); fwrite(handles.data.serialobj4,inttime_2); fwrite(handles.data.serialobj4,handles.data.startcmd); %PMT 5 % open serial port to PMT 5 handles.data.serialobj5=serial('COM105', 'BaudRate', 9600, 'Parity', 'none', 'InputBufferSize', 30000); fopen(handles.data.serialobj5); fwrite(handles.data.serialobj5,handles.data.stopcmd); fwrite(handles.data.serialobj5,measurenum); fwrite(handles.data.serialobj5,inttime_2); fwrite(handles.data.serialobj5,handles.data.startcmd); else % used to flush data and resend int time to get back more data % PMT 1

PAGE 211

211 flushinput(handles.data.serialobj1); fwrite(handles.data.serialobj1,inttime_1); fwrite(handles.data.serialobj1,handles.data.startcmd); %PMT 2 flushinput(handles.data.serialobj2); fwrite(handles.data.serialobj2,inttime_1); fwrite(handles.data.serialobj2,handles.data.startcmd); %PMT 3 flushinput(handles.data.serialobj3); fwrite(handles.data.serialobj3,inttime_1); fwrite(handles.data.serialobj3,handles.data.startcmd); %PMT 4 flushinput(handles.data.serialobj4); fwrite(handles.data.serialobj4,inttime_2); fwrite(handles.data.serialobj4,handles.data.startcmd); %PMT 5 flushinput(handles.data.serialobj5); fwrite(handles.data.serialobj5,inttime_2); fwrite(handles.data.serialobj5,handles.data.startcmd); end A=[]; B=[]; C=[]; D=[]; E=[]; handles.data.output1 = []; handles.data.output2 = []; handles.data.out1= []; handles.data.out2= []; handles.data.out3= []; handles.data.out4= []; handles.data.out5= []; handles.data.x1 = [handles.data.int_time_PMT1/1000]; handles.data.x2 = [handles.data.int_time_PMT2/1000]; i=1; j=1; k=1; m=1; n=1; total1=0; total2=0; % set PMT initial counts (total and integral) to 0 set(handles.PMT1_counts, 'string', num2str(0)); set(handles.PMT1_total, 'string', num2str(0)); set(handles.PMT2_counts, 'string', num2str(0));

PAGE 212

212 set(handles.PMT2_total, 'string', num2str(0)); guidata(hObject, handles); %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% while (get(handles.Start,'Userdata') == 1) if handles.data.serialobj3.BytesAvailable >= 5 C=fread(handles.data.serialobj3, 5); if C(1) == 65 handles.data.out3(k) = C(2)*256^3+C(3)*256^2+C(4)*256+C(5); handles.data.output1(3,k) = ((handles.data.out3(k)/(handles.data.int_time_PMT1/1000))/(1handles.data.dead_time_PMT3*(handles.data.out3(k)/(handles.data.int_time_PMT1 /1000))))*(handles.data.int_time_PMT1/1000); elseif C(1) == 66 set(handles.PMT1_counts,'string', 'Overflow'); drawnow Break else set(handles.PMT1_counts, 'string', 'Error!'); drawnow Break end k=k+1; d rawnow end %Gad Fiber if handles.data.serialobj2.BytesAvailable >= 5 B=fread(handles.data.serialobj2, 5); if B(1) == 65 handles.data.out2(j) = B(2)*256^3+B(3)*256^2+B(4)*256+B(5); handles.data.output1(2,j) = ((handles.data.out2(j)/(handles.data.int_time_PMT1/1000))/(1handles.data.dead_time_PMT2*(handles.data.out2(j)/(handles.data.int_time_PMT1 /1000))))*(handles.data.int_time_PMT1/1000); handles.data.output1(5,j)=(handles.data.output1(2,j)handles.data.output1(3,j))/(handles.data.int_time_PMT1/1000); elseif B(1) == 66 set(handles.PMT1_counts,'string', 'Overflow'); drawnow Break else set(handles.PMT1_counts, 'string', 'Error!'); drawnow Break end j=j+1; drawnow

PAGE 213

213 end %Plastic Fiber if handles.data.serialobj1.BytesAvailable >= 5 A=fread(handles.data.serialobj1, 5); if A(1) == 65 handles.data.upper=0.05*handles.data.center_ratio+handles.data.center_ratio; handles.data.lower=handles.data.center_ratio0.05*handles.data.center_ratio; handles.data.out1(i) = A(2)*256^3+A(3)*256^2+A(4)*256+A(5); handles.data.output1(1,i) = ((handles.data.out1(i)/(handles.data.int_time_PMT1/1000))/(1handles.data.dead_time_PMT1*(handles.data.out1(i)/(handles.data.int_time_PMT1 /1000))))*(handles.data.int_time_PMT1/1000); handles.data.output1(4,i)=(handles.data.output1(1,i)handles.data.ref_cal_fact_PMT1*handles.data.output1(3,i))/(handles.data.int_t ime_PMT1/1000); handles.data.output1(6,i)=handles.data.output1(4,i)/handles.data.output1(5,i) ; if handles.data.output1(1,i) > 15 ; handles.data.output1(4,i) > handles.data.cr_thresh_PMT1 ; handles.data.ouput1(6,i) < handles.data.upper ; handles.data.ouput1(6,i) > handles.data.lower ; total1=total1+(handles.data.int_time_PMT1/1000)*handles.data.ouput1(4,i)*hand les.data.prim_cal_fact_PMT1; end %%%%%%%%% OUTSIDE RATIO if handles.data.output1(1,i) > 15 ; handles.data.output1(4,i) > handles.data.cr_thresh_PMT1 ; handles.data.ouput1(6,i) < handles.data.lower ; total1=total1+(handles.data.int_time_PMT1/1000)*handles.data.ouput1(4,i)*hand les.data.prim_cal_fact_PMT1*handles.data.ratio_cal_fact*handles.data.output(6 ,i); end if handles.data.output1(1,i) > 15 ; handles.data.output1(4,i) > handles.data.cr_thresh_PMT1 ; handles.data.ouput1(6,i) > handles.data.upper ; total1=total1+(handles.data.int_time_PMT1/1000)*handles.data.ouput1(4,i)*hand les.data.prim_cal_fact_PMT1*handles.data.ratio_cal_fact*handles.data.output(6 ,i); end %%%%%%%%%%% LOW COUNT RATE if handles.data.output1(1,i) > 15 ; handles.data.output1(4,i) < handles.data.cr_thresh_PMT1

PAGE 214

214 total1=total1+(handles.data.int_time_PMT1/1000)*handles.data.ouput1(4,i)*hand les.data.sec_cal_fact_PMT1; end set(handles.PMT1_counts, 'string', num2str(handles.data.output1(6,i))); if i > 1 handles.data.x1(i) = handles.data.x1(i1)+handles.data.int_time_PMT1/1000; end plot(handles.axes1,handles.data.x1,handles.data.output1(6,:), 'b'); xlabel('Time (s)'); ylabel('Ratio'); elseif A(1) == 66 set(handles.PMT1_counts,'string', 'Overflow'); drawnow Break else set(handles.PMT1_counts, 'string', 'Error!'); drawnow Break end i=i+1; set(handles.PMT1_total, 'string', num2str(total1)); drawnow end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% if handles.data.serialobj5.BytesAvailable >= 5 E=fread(handles.data.serialobj5, 5); if E(1) == 65 handles.data.out5(n) = E(2)*256^3+E(3)*256^2+E(4)*256+E(5); handles.data.output2(2,n) = ((handles.data.out5(n)/(handles.data.int_time_PMT2/1000))/(1handles.data.dead_time_PMT5*(handles.data.out5(n)/(handles.data.int_time_PMT2 /1000))))*(handles.data.int_time_PMT2/1000); elseif E(1) == 66 set(handles.PMT2_counts,'string', 'Overflow'); drawnow Break else set(handles.PMT2_counts, 'string', 'Error!'); drawnow Break end n=n+1; drawnow end

PAGE 215

215 %%% Linear FOC fiber if handles.data.serialobj4.BytesAvailable >= 5 D=fread(handles.data.serialobj4, 5); if D(1) == 65 handles.data.out4(m) = D(2)*256^3+D(3)*256^2+D(4)*256+D(5); handles.data.output2(1,m) = ((handles.data.out4(m)/(handles.data.int_time_PMT2/1000))/(1handles.data.dead_time_PMT4*(handles.data.out4(m)/(handles.data.int_time_PMT2 /1000))))*(handles.data.int_time_PMT2/1000); handles.data.output2(3,m)=(handles.data.output2(1,m)handles.data.ref_cal_fact_PMT2*handles.data.output2(2,m))/(handles.data.int_t ime_PMT2/1000); %%%%%%%% DIRECT BEAM if handles.data.output2(1,m) > 15 ; handles.data.output2(3,m) > handles.data.crt_thresh_PMT2 ; total2=total2+handles.data.ouput2(3,m)*(handles.data.int_time_PMT2/1000)*hand les.data.prim_cal_fact_PMT2; end %%%%%%%% SCATTER BEAM if handles.data.output2(1,m) > 15 ; handles.data.output2(3,m) < handles.data.crt_thresh_PMT2 ; total2=total2+handles.data.ouput2(3,m)*(handles.data.int_time_PMT2/1000)*hand les.data.sec_cal_fact_PMT2; end set(handles.PMT2_counts, 'string', num2str(handles.data.output2(3,m))); if m > 1 handles.data.x2(m) = handles.data.x2(m1)+handles.data.int_time_PMT2/1000; end plot(handles.axes2,handles.data.x2,handles.data.output2(3,:), 'b'); xlabel('Time (s)'); ylabel('Count Rate'); elseif D(1) == 66 set(handles.PMT2_counts,'string', 'Overflow'); drawnow Break else set(handles.PMT2_counts, 'string', 'Error!'); drawnow Break end m=m+1; set(handles.PMT2_total, 'string', num2str(total2)); drawnow end

PAGE 216

216 end guidata(hObject, handles); % End of Start Button % Stop Button % --Executes on button press in Stop. function Stop_Callback(hObject, eventdata, handles) % hObject handle to Stop (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) set(handles.Start,'UserData',0); fwrite(handles.data.serialobj1,handles.data.stopcmd); fwrite(handles.data.serialobj2,handles.data.stopcmd); fwrite(handles.data.serialobj3,handles.data.stopcmd); fwrite(handles.data.serialobj4,handles.data.stopcmd); fwrite(handles.data.serialobj5,handles.data.stopcmd); guidata(hObject, handles); % Disconnect Button % --Executes on button press in Disconnect. function Disconnect_Callback(hObject, eventdata, handles) % hObject handle to Disconnect (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) handles.data.output1 = [0]; handles.data.x1 = [0]; handles.data.output2 = 0; handles.data.x2 = 0; % PMT 1 fclose(handles.data.serialobj1); axes(handles.axes1); plot(handles.data.x1,handles.data.output1); axis([0 1 0 1]) drawnow % PMT 2 fclose(handles.data.serialobj2); drawnow % PMT 3 fclose(handles.data.serialobj3); drawnow % PMT 4 fclose(handles.data.serialobj4); axes(handles.axes2);

PAGE 217

217 plot(handles.data.x2,handles.data.output2); axis([0 1 0 1]) drawnow % PMT 5 fclose(handles.data.serialobj5); drawnow handles.data.count = 1; guidata(hObject, handles); % File name function Filename_Callback(hObject, eventdata, handles) % hObject handle to Filename (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of Filename as text % str2double(get(hObject,'String')) returns contents of Filename as a double handles.data.Filename= get(hObject,'String'); guidata(hObject, handles); % --Executes during object creation, after setting all properties. function Filename_CreateFcn(hObject, eventdata, handles) % hObject handle to Filename (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end handles.data.Filename= get(hObject,'String'); guidata(hObject, handles); %Save to text file button Button % --Executes on button press in Filewrite. function Filewrite_Callback(hObject, eventdata, handles) % hObject handle to Filewrite (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) a=length(handles.data.output1); b=length(handles.data.output2); M=zeros(11,a); M(1,:)=handles.data.x1; M(2:7,:)=handles.data.output1;

PAGE 218

218 M(8,:)=handles.data.x2; M(9:11,:)=handles.data.output2; text_file=fopen(handles.data.Filename,'wt'); fprintf(text_file,'%s\ t %s\ t %s\ t %s\ t %s\ t %s\ t %s\ t %s\ t %s\ t %s\ t %s\ n', 'T(s)', 'PMT1', 'PMT2', 'PMT3', 'PSCR', 'GSCR', 'RATO', 'T(s)', 'PMT4', 'PMT5', LDCR'); fprintf(text_file,'%g\ t %g\ t %g\ t %g\ t %g\ t %g\ t %g\ t %g\ t %g\ t %g\ t %g\ n',M); guidata(hObject, handles); %%%%%%%% Display Boxes for total and integral counts for PMT 1 thru 3 %%%%%%%%% OUTPUT NUMBER BOXES %%%%%%%%%%%%%%%%%%%% function PMT1_counts_Callback(hObject, eventdata, handles) % hObject handle to PMT1_counts (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of PMT1_counts as text % str2double(get(hObject,'String')) returns contents of PMT1_counts as a double % --Executes during object creation, after setting all properties. function PMT1_counts_CreateFcn(hObject, eventdata, handles) % hObject handle to PMT1_counts (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end function PMT1_total_Callback(hObject, eventdata, handles) % hObject handle to PMT1_total (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of PMT1_total as text % str2double(get(hObject,'String')) returns contents of PMT1_total as a double % --Executes during object creation, after setting all properties. function PMT1_total_CreateFcn(hObject, eventdata, handles) % hObject handle to PMT1_total (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called

PAGE 219

219 % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end function PMT2_counts_Callback(hObject, eventdata, handles) % hObject handle to PMT2_counts (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of PMT2_counts as text % str2double(get(hObject,'String')) returns contents of PMT2_counts as a double % --Executes during object creation, after setting all properties. function PMT2_counts_CreateFcn(hObject, eventdata, handles) % hObject handle to PMT2_counts (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end function PMT2_total_Callback(hObject, eventdata, handles) % hObject handle to PMT2_total (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of PMT2_total as text % str2double(get(hObject,'String')) returns contents of PMT2_total as a double % --Executes during object creation, after setting all properties. function PMT2_total_CreateFcn(hObject, eventdata, handles) % hObject handle to PMT2_total (see GCBO) % eventdata reserved to be defined in a future version of MATLAB % handles empty handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor', 'white'); end

PAGE 220

220 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % END OF CODE

PAGE 221

221 APPENDIX B DERIVATION FOR WEIGHTING LINEAR FIBER OPTIC DOSIMETER MEASUREMENTS The elongated scintillating fiber, when properly calibrated, measures the absorbed dose along its axis By pairing two of these detectors orthogonal to one another, the dose in the imaging plane is more accurately measured, due to the nature of obtaining a one dimensional integral dose, combined with the additional information provided by the second detec tor. This type of dosimetry setup, as seen in Figure B 1, is very compatible with the UF anthropomorphic phantoms. Dosimeters are placed perpendicular to each other and intersect at the center of the organ. In this setup, the absorbed dose to the liver is being measured in one of the slices. Figure B 1. Dosimeter alignment for performing organ dosimetry. Liver Rt. Kidney Spleen Lt. Kidney

PAGE 222

222 Organs lying in the axial plane can be modeled as 2D geometric shapes such as a circle or ellipse, as demonstrated by early stylized phantoms. Over time, phantoms have evolved, and their organs have been rendered into complex shapes. The computational reference phantoms at the University of Florida (UF) are constructed using voxelized CT cross sectional images interpolated together and modified to fit the reference criteria. The UF anthropomorphic phantoms are derived from these hybrid phantoms and therefore contain structures that cannot be defined with a simple mathematical formula. However, it would not be unreasonable to treat these complex shapes as an ellipse or any other simple geometric shape defined by one or two parameters such as illustrated in Figure B 2 The ellipse shown below is defined by its radi a and b Figure B 2. Geometry of a n ellipse. An ellipse is also a good geometric approximation for an organ because of its smooth exterior contours that resemble the surfaces of structures in the human body. This is illustarted in Figure B 1 which has an equivalent organ (yellow lines) designated for each of the major organs. Another reason why an ellipse is a good mathematical model is because the ellipse is defined by two parameters. If a pair of detectors with sensitive lengths 2a and 2b were aligned along minor and major axis of a n organ, the average 2D dose could be estimated using the measurements from each detector and a b =

PAGE 223

223 applying a correctional factor to each measurement. The average organ dose is represented as a function of the two dosimeter readings, which are calibrated to t he average dose along their sensitive length corresponding to the dimensions of the organ (a and b). = ( , ) ( B 1 ) Looking at Figure B 3 the dosimeters are placed on the major and minor axis of the ellipse. The dosimeters measure the average dose along their respective axis. By definition the best estimate of the average organ dose will involve a component of each measured dose. The weighting of each component is based on the length of the major and minor axis of the organ. Therefore we can write the average organ dose as : = ( ) + ( ) ( B 2 ) Where f and g represent weighting factors that account for the contribution of each axis to the overall area of the ellipse. For example an ellipse similar to the shape of a circle should equally weight the major and minor axis component of measured absorbed dose. An ellipse shaped like a submarine should favor the dose measured along its major axis.

PAGE 224

224 Figure B 3. Dosimeters aligning to geometry of an ellipse. Now, assume that each detector can capture the dose profile at a distance dr from its axis as illustrated above. For an ellipsoid characterized along the x, y, and z axis by a b and c as shown below wher e a > b > c modeled as an organ where dosimeters 2a and 2b are used to capture organ dose data, the fractional dose capture by each detector is: = 2 ( ) 4 3 + 2 ( ) 4 3 ( B 3 ) However since we want the sum of f and g to equal 1we will divide by the common terms and after normalizing we get our final expression for the estimated absorbed dose : = + + + ( B 4 ) dr 2a 2b

PAGE 225

225 The above equation is only valid for an organ with a uniform density. Equation B 4 was used to weight the response of each dosimeter to obtai n an average organ/fetal dose.

PAGE 226

226 LIST OF REFERENCES 1 E. Lazarus, C. Debenedectis, D. North, P.K. Spencer, W.W. MayoSmith, "Utilization of imaging in pregnant patients: 10year review of 5270 examinations in 3285 patients -19972006," Radiology 251 517 524 (2009). 2 N ational Statistics, NHS Imaging and Radi ological Activity in England, National Statistics and National Health Service, London, England, 2013. 3 National Committee on Radiation Protection, "Exposure of the U.S. population from diagnostic medical radiation," NCRP Report No. 100 ( 1989). 4 Florida Department of Health, Chapter 64E 5, Florida Administrative C ode, X Ray in the Healing Arts, Florida Department of health,Tallahassee, FL, 2007 5 P.C. Shrimpton, M.C. Hillier, M.A. Lewis, M. Dunn, "National survey of doses from CT in the UK: 2003," Br J R adiol 79 968 980 (2006). 6 C.H. McCollough, M.R. Bruesewitz, J. J. M. Kofler, "CT dose reduction and dose management tools: overview of available options," Radiographics 26 503 512 (2006). 7 International Comission on Radiation Protection, "1990 recommendations of the International Commission on Radiological Protection," ICRP Publication 60 ( 1991). 8 J.M. Boone, "The trouble with CTDI_100," Med Phys 34, 1364 1371 (2007). 9 W. Huda, K.M. Ogden, M.R. Khorasani, "Converting doselength product to effective dose at CT," Radiology 248, 995 1003 (2008). 10 A. Abdullah, Z. Sun, N. Pongnapang, K.H. Ng, "Comparison of computed tomography dose reporting software," Radiat Prot Dosimetry 151, 153 157 (2012). 11 S. Mori, M. Endo, K. Nishizawa, T. Tsunoo, T. Aoyama, H. Fujiwara, K. Murase, "Enlarged longitudinal dose profiles in conebeam CT and the need for modified dosimetry," Med Phys 32, 10611069 (2005). 12 R.L. Dixon, J.M. Boone, "Dose equations for tube current modulation in CT scanning and the interpretation of the associated CTDIvol," Med Phys 40 111920 (2013). 13 M. Rosenstein, C.f. Devices, R. Health, Handbook of Selected Tissue Doses for Projections Common in Diagnostic Radiology (U.S. Department of Health and Human Services, Public Health Service, Food and Drug Administration, Center for Devices and Radiological Health, 1988). 14 International Comission on Radiation Protection, "The 2007 recommendations of the International Commission on Radiological Protection," I CRP Publication 103 ( 2007).

PAGE 227

227 15 D. Brenner, W. Huda, "Effective dose: a useful concept in diagnostic radiology," Radiat Prot Dosimetry 128, 503 508 (2008). 16 BEIR, Report No. BEIR VII, National Research Council Washington DC, 2005. 17 T.A. Jaffe, A.M. Neville, C. Anderson Evans, S. Long, C. Lowry, T.T. Yoshizumi, G. Toncheva, "Early first trimester fetal dose estimation method in a multivendor study of 16and 64 MDCT scanners and low dose imaging protocols," AJR. 193, 10191024 (2009). 18 C.H. McCollough, B.A. Schueler, T.D. Atwell, N.N. Braun, D. M. Regner, D.L. Brown, A.J. LeRoy, "Radiation exposure and pregnancy: when should we be concerned?," Radiographics 27 909917; discussion 917908 (2007). 19 L.M. Hurwitz, T. Yoshizumi, R.E. Reiman, P.C. Goodman, E.K. Paulson, D.P. Frush, G. Toncheva, G. N guyen, L. Barnes, "Radiation dose to the fetus from body MDCT during early gestation," AJR 186 871876 (2006). 20 J. Damilakis, A. Tzedakis, K. Perisinakis, A.E. Papadakis, "A method of estimating conceptus doses resulting from multidetector CT examinatio ns during all stages of gestation," Med Phys 37, 64116420 (2010). 21 E. Angel, C.V. Wellnitz, M.M. Goodsitt, N. Yaghmai, J.J. DeMarco, C.H. Cagnon, J.W. Sayre, D.D. Cody, D.M. Stevens, A.N. Primak, C.H. McCollough, M.F. McNittGray, "Radiation dose to the fetus for pregnant patients undergoing multidetector CT imaging: Monte Carlo simulations estimating fetal dose for a range of gestational age and patient size," Radiology 249, 220 227 (2008). 22 L.B. Russell, W.L. Russell, "An analysis of the changing radiation response of the developing mouse embryo," Journal of cellular physiology. Supplement 43 103 149 (1954). 23 B.M. Carlson, Human Embryology and Development Biology (Mosby Year Book Inc., St. Louis, 1994). 24 C. Streffer, W.U. Muller, "Malformations after radiation exposure of preimplantation stages," Int J Dev Biol 40, 355360 (1996). 25 I nternation Commission on Radiation Protection, "Biological effects after prenatal irradiation (embryo and fetus)," ICRP Publication 90 (2 003). 26 E.J. Hall, Radiobiology for the Radiobiologist 4th ed. (Lippincott, Philadelphia, 1994). 27 A.S. Dekaban, "Abnormalities in children exposed to x radiation during various stages of gestation: tentative timetable of radiation injury to the human fetus. I," J Nucl Med 9 471477 (1968).

PAGE 228

228 28 UNSCEAR, Sources and Effects of Ionizing Radiation Report to General Assembly, 1993. 29 UNSCEAR, Genetic and somantic Effects of Ionizing Radiation, Report to General Assembly, 1986. 30 M. Otake, W.J. Schull, S. Lee, "Thre shold for radiationrelated severe mental retardation in prenatally exposed A bomb survivors: a reanalysis," Int J Radiat Biol 70 755763 (1996). 31 International Commission on Radiation Protection, "Development effects of irradiation on the brain of the embryo and fetus," ICRP Publication 49 ( 1986). 32 E.G. Knox, T.T. Odell, Jr., E.P. Sniffen, "Prenatal irradiation and childhood cancer," J. Soc.Radiol. Prot. 7 177 189 (1987). 33 A.C. Upton, T.T. Odell, Jr., E.P. Sniffen, "Influence of age at time of irr adiation on induction of leukemia and ovarian tumors in RF mice," Proceedings of the Society for Experimental Biology and Medicine. Society for Experimental Biology and Medicine (New York, N.Y.) 104, 769772 (1960). 34 R.R. Monson, B. MacMahon, "Prenatal x ray exposure and cancer in children," in Radiation Carcinogenesis: Epidemiology and Biological Significance [ edited by ] J.F. Fraumeni (Raven Press, New York, 1984), pp. 99105. 35 D. Hewitt, B. Sanders, A. Stewart, "Oxford Survey of Childhood Cancers: progress report. IV. Reliability of data reported by case and control mothers," Monthly bulletin of the Ministry of Health and the Public Health Laboratory Service 25, 80 85 (1966). 36 R. Wakeford, M.P. Little, "Childhood cancer after low level intrauterine exposure to radiation," J Soc Radiol Prot 22 A123127 (2002). 37 A. Berrington de Gonzalez, A. Iulian Apostoaei, L.H. Veiga, P. Rajaraman, B.A. Thomas, F. Owen Hoffman, E. Gilbert, C. Land, "RadRAT: a radiation risk assessment tool for lifetime cancer risk projection," Soc Radiol Prot 32 205 222 (2012). 38 D.C. Kocher, A.I. Apostoaei, R.W. Henshaw, F.O. Hoffman, M.K. Schubauer Berigan, D.O. Stancescu, B.A. Thomas, J.R. Trabalka, E.S. Gilbert, C.E. Land, "Interactive RadioEpidemiological Program (IREP): a w ebbased tool for estimating probability of causation/assigned share of radiogenic cancers," Health Phys 95, 119 147 (2008). 39 H ealth Physics Society, "Uncerainty in Risk Assessment," Position Statements of the Health physics Society, 2013. 40 J. Bergoni, L. Tribondeau, "De quelques rsultats de la radiotherapie et essai de fixation d'une technique rationnelle," Comptes Rendus des Sances de l'Acadmie des Sciences 149 983 985 (1906).

PAGE 229

229 41 ACOG Committee Opinion. Number 299, September 2004 (repl aces No. 158, September 1995). Guidelines for diagnostic imaging during pregnancy," Am J Obstet and Gynecol 104 647651 (2004). 42 B.F. Wall, J.R. Meara, C.R. Muirhead, R.F. Bury, M. Murray, "Protection of Pregnant Patients during Diagnostic Medical Expos ures to Ionising Radiation," Health Protection Agency and The Royal College of Radiologists and the College of Radiographers, 2009. 43 International Comission of Radiation Protection, "Pregnancy and medical radiation," ICRP Publication 84 ( 2000). 44 G.V. B ochicchio, L.M. Napolitano, J. Haan, H. Champion, T. Scalea, "Incidental pregnancy in trauma patients," J Am Col l of Surg 192, 566569 (2001). 45 B.L. Justus, P. Falkenstein, A.L. Huston, M.C. Plazas, H. Ning, R.W. Miller, "Gated fiber optic coupled detect or for in vivo real time radiation dosimetry," Appl Opt 43 16631668 (2004). 46 C.J. Marckmann, M.C. Aznar, C.E. Andersen, L. Botter Jensen, "Influence of the stem effect on radioluminescence signals from optical fibre Al2O3:C dosemeters," Radiat Prot Dos imetry 119, 363 367 (2006). 47 L. Archambault, A.S. Beddar, L. Gingras, R. Roy, L. Beaulieu, "Measurement accuracy and Cerenkov removal for high performance, high spatial resolution scintillation dosimetry," Med Phys 33, 128135 (2006). 48 F. Lacroix, L. A rchambault, L. Gingras, M. Guillot, A.S. Beddar, L. Beaulieu, "Clinical prototype of a plastic water equivalent scintillating fiber dosimeter array for QA applications," Med Phys 35, 36823690 (2008). 49 A.S. Beddar, T.R. Mackie, F.H. Attix, "Cerenkov ligh t generated in optical fibers and other light pipes irradiate by electron beams," Phys Med Biol 37 925 935 (1992). 50 A.S. Beddar, T.M. Briere, F.A. Mourtada, O.N. Vassiliev, H.H. Liu, R. Mohan, "Monte Carlo calculations of the absorbed dose and energy dependence of plastic scintillators," Med Phys 32, 12651269 (2005). 51 D.E. Hyer, University of Florida, 2010. 52 A.S. Beddar, "Plastic scintillation dosimetry and its application to radiotherapy," Radiat Meas 41, S124S 133 (2006). 53 A.S. Beddar, T.R. Mackie, F.H. Attix, "Water equivalent plastic scintillation detectors for high energy beam dosimetry: II. Properties and measurements," Phys Med Biol 37, 19011913 (1992).

PAGE 230

230 54 D. Letourneau, J. Pouliot, R. Roy, "Miniature scintillating detector for small field radiation therapy," Med Phys 26, 25552561 (1999). 55 A.S. Beddar, K.J. Kinsella, A. Ikhlef, C.H. Sibata, "A miniature scintillator fiberoptic PMT; detector system for the dosimetry of small fields in stereotactic radiosurgery," IEEE Trans Nucl Sci 48 924 928 (2001). 56 L. Archambault, A.S. Beddar, L. Gingras, F. Lacroix, R. Roy, L. Beaulieu, "Water equivalent dosimeter array for small field external beam radiotherapy," Med Phys 34, 15831592 (2007). 57 M. Guillot, L. Gingras, L. Archambault, S. Beddar, L. Beaulieu, "Spectral method for the correction of the Cerenkov light effect in plastic scintillation detectors: a comparison study of calibration procedures and validation in Cerenkov light dominated situations," Med Phys 38 21402150 (2011). 58 A.S. Beddar, T.R. Mackie, F.H. Attix, "Water equivalent plastic scintillation detectors for high energy beam dosimetry: I. Physical characteristics and theoretical consideration," Phys Med Biol 37 18831900 (1992). 59 D. Fluhs, M. Heintz, F. Indenkampen, C. Wieczorek, H. Kolanoski, U. Quast, "Direct reading measurement of absorbed dose with plastic scintillators The general concept and applications to ophthalmic plaque dosimetry, Med Phys 23, 427 434 (1996). 60 P.L. Mattern, L.M. Watkins, C.D. Skoog, E.H. Barsis, "Absorption induced in optical waveguides by pulsed electrons as a function of temperature, low dose rate gamma and beta rays, and 14 MeV nuetrons," IEEE Trans Nucl Sci NS 22 2468 (1975). 61 P.L. Mattern, L.M. Watkins, C.D. Skoog, J.R. Brandon, E.H. Barsis, "The effects of radiation on the absorption and luminescence of fiber optic waveguides and materials," IEEE Trans Nucl Sci NS 21, 81 (1974). 62 S. Girard, J. Keurinck Charre, M. Vi, "Gamma rays and pulsed X ray radiation responses of nitrogen, germanium doped and pure silica core optical fibers," Nucl Instrum Meth B 215 187 195 (2004). 63 R.L. Clough, K.T. Gillen, G.M. Malone, J.S. Wallace, "Color formation in irradiated polymers," Radiat Phys Chem 48 583 594 (1996). 64 J.G. Titchmarsh, "Signal level dependence in the recovery rate of irradiated plastic clad silica fibre," Electron Libr 15, 111 112 (19 79). 65 E.J. Friebele, C.G. Askins, M.E. Gingerich, K.J. Long, "Optical fiber waveguides in radiation environments, II," Nucl Instrum Meth B 1 355 369 (1984).

PAGE 231

231 66 W. Busjan, K. Wick, T. Zoufal, "On the behaviour of plastic scintillators during an experiment in a high dose rate environment," Nucl Instrum Meth B 151 434437 (1999). 67 K.T. Gillen, J.S. Wallace, R.L. Clough, "Doserate dependence of the radiationinduced discoloration of polystyrene," Radiat Phys Chem 41, 101 113 (1993). 68 J.E. Golob, P.B. Lyons, L.D. Looney, "Transient Radiation Effects in Low Loss Optical Waveguides," IEEE Trans Nucl Sci 24, 21642168 (1977). 69 P.A. Jursinic, "Dependence of diode sensitivity on the pulse rate of delivered radiation," Med Phys 40 021720 ( 2013). 70 R.D. Evans, "Effects of finite resolving time on counting instruments," in The Atomic Nucleus (McGraw Hill, New York, 1955), pp. 785794. 71 V. Wegst, A. Ashare, J.A. Correia, J.C. Ehrhardt, A.L. Forsaith, L.S. Graham, W.A. Roventine, J.A. Sorens on, D.A. Weber, R.L. Witcofski, R.E. Zimmerman, "AAPM Report No. 6 Scintillation camera acceptance testing and performance evaluation," (American Institute of Physics, 1980). 72 H. Kume, K. Koyama, K. Nakatsugawa, S. Suzuki, D. Fatlowitz, "Ultrafast micr ochannel plate photomultipliers," Appl. Opt. 27 1988). 73 H amamstsu, Hamamatsu Photonics, 2007. 74 S.R. Cherry, J.A. Sorenson, Physics in Nuclear Medicine 4th ed. (Sanders, Philadelphia, 2012). 75 T. Woldeselassie, "Modeling of scintillation camera systems ," Med Phys 26, 1375 1381 (1999). 76 G.F. Knoll, Radiation Detection and Measurement 3rd ed. (John Wiley & Sons, Hoboken, 2000). 77 H. Daams, "The apparent dead time of a geiger counter, measuring an intensity of radiation which varies periodically," in Ap pl Sci Res 3 29 34 (1954) 78 D.E. Hyer, R.F. Fisher, D.E. Hintenlang, "Characterization of a water equivalent fiber optic coupled dosimeter for use in diagnostic radiology," Med Phys 36, 17111716 ( 2009). 79 G. Ayotte, L. Archambault, L. Gingras, F. Lacroix, A.S. Beddar, L. Beaulieu, "Surface preparation and coupling in plastic scintillator dosimetry," Med Phys 33, 35193525 (2006).

PAGE 232

232 80 J. Elsey, D.R. McKenzie, J. Lambert, N. Suchowerska, S.L. Law, S.C. Fleming "Optimal coupling of light from a cylindrical scintillator into an optical fiber," Appl Opt 46, 397404 (2007). 81 L. Torrisi, "Plastic scintillator investigations for relative dosimetry in protontherapy," Nucl Instrum Meth B 170 523530 (2000). 82 L.L .W. Wang, L.A. Perles, L. Archambault, N. Sahoo, D. Mirkovic, S. Beddar, "Determination of the quenching correction factors for plastic scintillation detectors in therapeutic highenergy proton beams," Phys in Med Biol 57 77677775 (2012). 83 M. Ghergherehchi, H. Afarideh, M. Ghannadi, A. Mohammadzadeh, G.R. Aslani, B. Boghrati, "Proton beam dosimetry: a comparison between a plastic scintillator, ionization chamber and Faraday cup," Journal of radiation research 51, 423 430 (2010). 84 F.H. Attix, Introduct ion to Radiological Physics and Radiation Dosimetry (Wiley VCH, Hoboken, NJ, 1986). 85 F. Pernicka, I.D. McLean, I.A.E. Agency, Dosimetry in Diagnostic Radiology: An International Code of Practice. (International Atomic Energy Agency, 2007). 86 L.A. DeWe rd, L.K. Wagner, "Characteristics of radiation detectors for diagnostic radiology," Int Appl R adiat I so 50 125 136 (1999). 87 T. Kron, L. Duggan, T. Smith, A. Rosenfeld, M. Butson, G. Kaplan, S. Howlett, K. Hyodo, "Dose response of various radiation detec tors to synchrotron radiation," Phys Med Biol 43, 32353259 (1998). 88 S.L. Dong, T.C. Chu, G.Y. Lan, T.H. Wu, Y.C. Lin, J.S. Lee, "Characterization of high sensitivity metal oxide semiconductor field effect transistor dosimeters system and LiF:Mg,Cu,P the rmoluminescence dosimeters for use in diagnostic radiology," Int Appl R adiat I so 57 883891 (2002). 89 J. Zoetelief, P. Christensen, H.W. Julius, W. Leitz, B.M. Moores, E.C.R. DirectorateGeneral, O.f.O.P.o.t.E. Communities, W. Panzer, B.F. Wall, European Commission, Recommendations for patient dosimetry in diagnostic radiology using TLD (EUR, 2000). 90 J.V. Valiyaparambil, S.M. Mallya, "Characterization of an optically stimulated dosimeter for dentomaxillofacial dosimetry," Oral Surg O ral Med O ral Path O ral Radiol Endod 112, 793 797 (2011). 91 E.G. Yukihara, S.W. McKeever, Optically stimulaed luminescence: fundamentals and applications (John Wiley & Sons Ltd, United Kingdom, 2011). 92 R.M. Al Senan, M.R. Hatab, "Characteristics of an OSLD in the diagnost ic energy range," Med Phys 38 43964405 (2011).

PAGE 233

233 93 A. Endo, T. Katoh, I. Kobayashi, R. Joshi, J. Sur, T. Okano, "Characterization of optically stimulated luminescence dosemeters to measure organ doses in diagnostic radiology," D MFR 41, 211216 (2012). 94 A.K. Jones, F.D. Pazik, D.E. Hintenlang, W.E. Bolch, "MOSFET dosimeter depthdose measurements in heterogeneous tissueequivalent phantoms at diagnostic x ray energies," Med Phys 32, 32093213 (2005). 95 M.C. Lavallee, L. Gingras, L. Beaulieu, "Energy and integrated dose dependence of MOSFET dosimeter sensitivity for irradiation energies between 30 kV and 60Co," Med Phys 33, 3683 3689 (2006). 96 L.A. Benevides, A.L. Huston, B.L. Justus, P. Falkenstein, L.F. Brateman, D.E. Hintenlang, "Characterization of a fiber optic coupled radioluminescent detector for application in the mammography energy range," Med Phys 34, 22202227 (2007). 97 A.K. Jones, D. Hintenlang, "Potential clinical utility of a fibre optic coupled dosemeter for dose measurements in diagnostic radiology," Radiat Prot Dosimetry 132, 80 87 (2008). 98 M. Moloney, University of Florida, 2008. 99 F. Lessard, L. Archambault, M. Plamondon, P. Despres, F. Therriault Proulx, S. Beddar, L. Beaulieu, "Validating plastic scintillation detectors for photon dosimetry in the radiologic energy range," Med Phys 39, 53085316 (2012). 100 L. Lavoie, M. Ghita, L. Brateman, M. Arreola, "Characterization of a commercially available, optically stimulated luminescent dosimetry system for use in computed tomography," Health Phys 101, 299310 (2011). 101 J.F. Williamson, J.F. Dempsey, A.S. Kirov, J.I. Monroe, W.R. Binns, H. Hedtjarn, "Plastic scintillator response to low energy photons," Phys Med Biol 44, 857871 (1999). 102 E. Brannen, G.L. Olde, "The response of organic scintillators to electron en ergy deposited in them," Radiat Res 16, 1 6 (1962). 103 A.J. Meyerott, P.C. Fisher, D.T. Roethig, "Plastic Scintillator Response to 1 10 keV Photons," Rev Sci Instrum 35 669672 (1964). 104 A.M. Frelin, J.M. Fontbonne, G. Ban, J. Colin, M. Labalme, "Comparative Study of Plastic Scintillators for Dosimetric Applications," IEEE Trans Nucl Sci 55, 27492756 (2008). 105 L. Dixon, D. M. Bakalyar, J. M. Boone, D. D. Cody, R. Fahrig, D. A. Jaffray, I. S. Kyprianou, C. H. McCollough, M. F. McNitt Gray, H. T. Morin, K. D. Nakonechny, T. J. Payne, R. J. Pizzutiello, B. T. Schmidt, A. J. Seibert, W. E. Simon, T. W. Slowery, S. H Stern, P. Sunde, T. L. Toth, and S. Vastagh, "Comprehensive methodology for

PAGE 234

234 the evaluation of radiation dose in x ray computed tomography. Report of AAPM Task Group 111: The Future of CT Dosimetry," American Association of Physicists in Medicine, 2010. 106 C. M. Ma, C.W. Coffey, L.A. DeWerd, C. Liu, R. Nath, S.M. Seltzer, J.P. Seuntjens, "AAPM protocol for 40300 kV x ray beam dosimetry in radiotherapy and radiobiology," Med Phys 28, 868 893 (2001). 107 A.S. Beddar, S. Law, N. Suchowerska, T.R. Mackie, Plastic scintillation dosimetry: optimization of light collection efficiency," Phys Med Biol 48, 11411152 (2003). 108 M.J. Berger, J.H. Hubbell, S.M. Seltzer, J. Chang, J.S. Coursey, R. Sukumar, D.S. Zucker, K. Olsen, "XCOM: Photon Cross Section Database (version 1.5).", Vol. 2013 (National Institute of Standards and Technology, Gaithersburg, MD, 2010). 109 M.J. Berger, J. Chang, J.S. Coursey, M.A. Zucker, J. Chang, "ESTAR, PSTAR, and ASTAR: Computer Programs for Calculating Stopping Power and Range Tables for Electrons, Protons, and Helium Ions (version 1.2.3)," Vol. 2013 (National Institute of Standards and Technology, Gaithersburg, MD, 2005). 110 A.C. Turner, D. Zhang, H.J. Kim, J.J. DeMarco, C.H. Cagnon, E. Angel, D.D. Cody, D.M. Stevens, A.N. Primak, C .H. McCollough, M.F. McNitt Gray, "A method to generate equivalent energy spectra and filtration models based on measurement for multidetector CT Monte Carlo dosimetry simulations," Med Phys 36, 21542164 (2009). 111 M. Ghita, University of Florida, 2009. 112 J.H. Siewerdsen, A.M. Waese, D.J. Moseley, S. Richard, D.A. Jaffray, "Spektr: a computational tool for x ray spectral analysis and imaging system optimization," Med Phys 31, 3057 3067 (2004). 113 J.M. Boone, J.A. Seibert, "An accurate method for comput er generating tungsten anode x ray spectra from 30 to 140 kV," Med Phys 24, 16611670 (1997). 114 J.B. Birks, F.D. Brooks, "Scintillation Response of Anthracene to 630 keV Photoelectrons," Proc Phys Soc London Sec B 69, 721 (1956). 115 R.L. Craun, D.L. Sm ith, "Analysis of response data for several organic scintillators," Nucl Instrum Meth 80 239244 (1970). 116 M. Hirschberg, R. Beckmann, U. Brandenburg, H. Bruckmann, K. Wick, "Precise measurement of Birks kB parameter in plastic scintillators," IEEE Tran s Nucl Sci 39 511514 (1992). 117 V.I. Tretyak, "Semi empirical calculation of quenching factors for ions in scintillators," Astropart. Phys. 33 4053 (2010).

PAGE 235

235 118 K. Jordan, "Review of recent advances in non 3D dosimeters," Journal of Physics: Conference Series 444, 012009 (2013). 119 J.B. Sessions, J.N. Roshau, M.A. Tressler, D.E. Hintenlang, M.M. Arreola, J.L. Williams, L.G. Bouchet, W.E. Bolch, "Comparisons of point and average organ dose within an anthropomorphic physical phantom and a computational m odel of the newborn patient," Med Phys 29, 10801089 (2002). 120 J.A. Bauhs, T.J. Vrieze, A.N. Primak, M.R. Bruesewitz, C.H. McCollough, "CT dosimetry: comparison of measurement techniques and devices," Radiographics 28 245253 (2008). 121 D.E. Hyer, C.F. Serago, S. Kim, J.G. Li, D.E. Hintenlang, "An organ and effective dose study of XVI and OBI conebeam CT systems," J App Clin Med Phys 11, 181 197 (2010). 122 A. Cozannet, M. Treheux, "Skew rays in optical fibers," Appl. Opt. 14 13451350 (1975). 123 R.F Fisher, Doctoral dissertation, University of Florida, 2010. 124 M.W. Kan, L.H. Leung, W. Wong, N. Lam, "Radiation dose from cone beam computed tomography for imageguided radiation therapy," Int J Radiat Oncol Biol Phys 70, 272 279 (2008). 125 D.P. Frush, T. Yoshizumi, "Conventional and CT angiography in children: dosimetry and dose comparisons," Pediatr Radiol 36 154 158 (2006). 126 O. Luz, M. Buchgeister, M. Klabunde, T. Trabold, A.F. Kopp, C.D. Claussen, M. Heuschmid, "Evaluation of dose exposure in 64slice CT colonography," Eur J Radiol 17, 26162621 (2007). 127 X.G. Xu, K.F. Eckerman, Handbook of Anatomical Models for Radiation Dosimetry (Taylor & Francis, 2009). 128 J.F. Winslow, D.E. Hyer, R.F. Fisher, C.J. Tien, D.E. Hintenlang, "Cons truction of anthropomorphic phantoms for use in dosimetry studies," J App Clin Med Phys 10, 195204 (2009). 129 C. Lee, D. Lodwick, D. Hasenauer, J.L. Williams, C. Lee, W.E. Bolch, "Hybrid computational phantoms of the male and female newborn patient: NURB S based wholebody models," Phys Med Biol 52 33093333 (2007). 130 C. Lee, D. Lodwick, J. Hurtado, D. Pafundi, J.L. Williams, W.E. Bolch, "The UF family of reference hybrid phantoms for computational radiation dosimetry," Phys Med Biol 55, 339363 (2010).

PAGE 236

236 131 J.L. Hurtado, C. Lee, D. Lodwick, T. Goede, J.L. Williams, W.E. Bolch, "Hybrid computational phantoms representing the reference adult male and adult female: construction and applications for retrospective dosimetry," Health Phys 102, 292 304 (2012). 132 International Comission of Radiation Protection, "Basic Anatomical and Physiological Data for Use in Radiological Protection Reference Values," ICRP Publication 89 ( 2002). 133 E.K. Osei, C.J. Kotre, "Equivalent dose to the fetus from occupational expos ure of pregnant staff in diagnostic radiology," Br J Radiol 74 629 637 (2001). 134 S.T. Schindera, R.C. Nelson, E.R. Lee, D.M. Delong, G. Ngyen, G. Toncheva, T.T. Yoshizumi, "Abdominal multislice CT for obese patients: effect on image quality and radiation dose in a phantom study," Acad R adiol 14, 486494 (2007). 135 S.T. Schindera, R.C. Nelson, T.L. Toth, G.T. Nguyen, G.I. Toncheva, D.M. DeLong, T.T. Yoshizumi, "Effect of patient size on radiation dose for abdominal MDCT with automatic tube current modulation: phantom study," AJR 190 W100 105 (2008). 136 D.J. Long, Doctoral dissertation, University of Florida, 2013. 137 M.R. Maynard, Doctoral dissertation, University of Florida, 2013. 138 F.G. Cunningham, J.W. Williams, Williams obstetrics 23rd ed. (McGr aw Hill Medical, New York, 2010). 139 A.H. DeCherney, L. Nathan, Current obstetric & gynecologic diagnosis & treatment 9th ed. / [edited by] Alan H. DeCherney, Lauren Nathan. ed. (Lange Medical Books/McGraw Hill, New York, 2003). 140 S.K. Thompson, S.M. Goldman, K.B. Shah, P.C. Chen, L.K. Wagner, F.M. Corl, B.K. Raval, S. Sheth, E.K. Fishman, C.M. Sandler, "Acute nontraumatic maternal illnesses in pregnancy: imaging approaches," Emerg Radiol 11, 199 212 (2005). 141 A.A. Al Mulhim, "Acute appendicitis in pregnancy. A review of 52 cases," Int Surg 81, 295297 (1996). 142 R.E. Andersson, M. Lambe, "Incidence of appendicitis during pregnancy," Int J Epidemiol 30 12811285 (2001). 143 M. Tracey, H.S. Fletcher, "Appendicitis in pregnancy," Amer Surg 66 555 55 9; discussion 559560 (2000). 144 M.K. Shetty, "Abdominal Computed Tomography During Pregnancy: A Review of Indications and Fetal Radiation Exposure Issues," Seminars in Ultrasound, CT and MRI 31, 3 7 (2010).

PAGE 237

237 145 M. Remy Jardin, M. Pistolesi, L.R. Goodman, W.B. Gefter, A. Gottschalk, J.R. Mayo, H.D. Sostman, ""Management of suspected acute pulmonary embolism in the era of CT angiography: a statement from the Fleischner Society," Radiology 245, 315 329 (2007). 146 C.J. Berg, H.K. Atrash, L.M. Koonin, M. Tuck er, "Pregnancy related mortality in the United Stat es, 1987 1990," Am J Obstet and G ynecol 88, 161 167 (1996). 147 H.T. Winer Muram, J.M. Boone, H.L. Brown, S.G. Jennings, W.C. Mabie, G.T. Lombardo, "Pulmonary embolism in pregnant patients: fetal radiation dose with helical CT," Radiology 224, 487 492 (2002). 148 N.S. Long, Masters thesis, University of Florida, 2013. 149 Mitsubishi Rayon Co. Ltd: Eksa Optical Fiber Division, "Specification Sheet: CK 20 (Mitsubishi Rayon Co., Ltd., Tokyo, Japan, 2001).

PAGE 238

238 BIOGRAPHICAL SKETCH Matthew Robert Hoerner was born in Pittsburgh, Pennsylvania, to Karen Flory and Richard Hoerner. He is one of two children, along with his younger brother, Michael. He graduated from Unionville High School in Kennett Square, Pennsylvania in 2005. Four years later he graduated from The Pennsylvania State University with a Bachelor of Science degree in nuclear engineering. In 2009, Matthew enrolled into the medical physics graduate program at the University of Florida where he com pleted his Master of Science and Doctor of Philosophy. After graduation, Matthew will begin his two year imaging physi cs residency at UF Health Shands Hospital in Gainesville, FL.